Previously, all fields were unsigned, with only a trailing "ago" to
indicate negative intervals. Now, ISO format does not use "ago", and
and the traditional PostgreSQL format has the first numeric field unsigned
with "ago" supporting that field. So "1 month - 2 days ago" is two days
less than a month in the past.
Fix interval arithmetic across daylight savings time boundaries.
Previously, most math across boundaries introduced a one hour offset.
Allow some date/time functions to return NULL if called with NULL args.
Implement functions for AT TIME ZONE support.
Support "SAT" as an Australian time zone if USE_AUSTRALIAN_RULES
is defined.
any available string type. Previously, all candidate choices must have
fallen within the same "type category" for PostgreSQL to be willing to
choose any of them.
Need to apply the same fixup to operator type resolution.
confused in src/interfaces/libpq for some reason. Seemingly, different
GNU make versions have several mutually conflicting problems with implicit
rule chains. Words are not sufficient...
target files in implicit rule chains. That might have been a cool idea
but it seems to be too buggy to work, as it caused spurious recompiles in
several places.
included, and then include <strings.h> if so. Several systems already
needed <strings.h> anyway. Some new systems that claim to conform to the
Unix 9x "standard" do not declare str[n]casemp() in string.h, and C99
compilers will not like that.
Bruce Hartzler <bruceh@mail.utexas.edu>. It contains shared library
support, regression test map, and the usual template files. The dynamic
loader is missing, the spin lock code apparently doesn't assemble due to
syntax problems, and semaphores are to be hoped for from Apple.
position() and substring() functions, so that it works transparently for
bit types as well. Alias the text functions appropriately.
Add position() for bit types.
Add new constant node T_BitString that represents literals of the form
B'1001 and pass those to zpbit type.
code conversion between Unicode and other encodings. Note that
this option requires --enable-multibyte also.
The reason why this is optional is that the feature requires huge
mapping tables and I don't think every user need the feature.
equivalent.
In linux.h there were some #undef HAVE_INT_TIMEZONE, which are useless
because HAVE_TM_ZONE overrides it anyway, and messing with configure
results isn't cool.
subqueries. It passes the normal 'runcheck' tests, and I've tried
a few simple things like
select 1 as foo union (((((select 2))))) order by foo;
There are a few things that it doesn't do that have been talked
about here at least a little:
1) It doesn't allow things like "IN(((select 1)))" -- the select
here has to be at the top level. This is not new.
2) It does NOT preserve the odd syntax I found when I started looking
at this, where a SELECT statement could begin with parentheses. Thus,
(SELECT a from foo) order by a;
fails.
I have preserved the ability, used in the regression tests, to
have a single select statement in what appears to be a RuleActionMulti
(but wasn't -- the parens were part of select_clause syntax).
In my version, this is a special form.
This may cause some discussion: I have differentiated the two kinds
of RuleActionMulti. Perhaps nobody knew there were two kinds, because
I don't think the second form appears in the regression tests. This
one uses square brackets instead of parentheses, but originally was
otherwise the same as the one in parentheses. In this version of
gram.y, the square bracket form treats SELECT statements the same
as the other allowed statements. As discussed before on this list,
psql cannot make sense out of the results of such a thing, but an
application might. And I have designs on just such an application.
++ kevin o'gorman
path into executables and shared libraries (-rpath or -R for most). Can be
disabled with --disable-rpath, since some binary packaging standards do not
like this option.
on pghackers. Arrange for the sort ordering of general INET values
to be network part as major sort key, host part as minor sort key.
I did not force an initdb for this change, but anyone who's running
indexes on general INET values may need to recreate those indexes.
ExecutorRun. This allows LIMIT to work in a view. Also, LIMIT in a
cursor declaration will behave in a reasonable fashion, whereas before
it was overridden by the FETCH count.
MULTIBYTE support is not compiled (you just can't set them to anything
but SQL_ASCII). This should reduce interoperability problems between
MB-enabled clients and non-MB-enabled servers.
support is not present. This allows a non-MB server to load a pg_dumpall
script produced by an MB-enabled server, so long as only ASCII encoding
was used.
the -l options. (This was not the case when using the OpenSSL or Kerberos
options.) Also make sure that shared library links get to see all the -L
options. Get Kerberos 5 support to compile on Redhat 7.0. Add OpenSSL and
-lsocket (if used/found) to libpq link.
I modified the current ODBC driver for
* referential integrity error reporting,
* SELECT in transactions and
* disabling autocommit.
I tested these changes with Borland C++ Builder -> ODBCExpress ->
WinODBC driver (DLL) -> Postgres 7.0beta1 and Borland C++ Builder -> BDE ->
WinODBC driver (DLL) -> Postgres 7.0beta1. The patch is based on snapshot of
22th April (I don't think that someone has modified it since that: Byron
hasn't gave any sign of living for about a month and I didn't find any
comments about the ODBC driver on the list).
'AbortTransaction and not in in-progress state' when client disconnects
just after an error. Notice seems pretty harmless, so I'm not going
to worry about back-patching this into 7.0.* ...
as full as possible, seems better to use a tuple size around BLCKSZ/4
so that less space is wasted when a LO tuple is updated. Also, this
lets us use a logical page size that's an exact power of two, avoiding
partial-page writes when client is sending us stuff in power-of-2
buffer chunks.
kibitzing from Tom Lane. Large objects are now all stored in a single
system relation "pg_largeobject" --- no more xinv or xinx files, no more
relkind 'l'. This should offer substantial performance improvement for
large numbers of LOs, since there won't be directory bloat anymore.
It'll also fix problems like running out of locktable space when you
access thousands of LOs in one transaction.
Also clean up cruft in read/write routines. LOs with "holes" in them
(never-written byte ranges) now work just like Unix files with holes do:
a hole reads as zeroes but doesn't occupy storage space.
INITDB forced!
source, due to addition of header overhead), store it as plain data
rather than pseudo-compressed data. This saves a few microseconds when
reading it out, but much more importantly guarantees that the toaster
won't actually expand tuples that contain incompressible data. That's
essential to avoid 'Tuple too big' failures with large objects.
particular, allow linking with arbitrary commands rather than only $(AR) or
$(LD), and treat C++ without hacks.
Add option to disable shared libraries. This takes the place of the
BSD_SHLIB variable. The regression test driver ignores the plpgsql test
if there are no shared libraries available.
from bufmgr - it would be nice to have separate hash in smgr
for node <--> fd mappings, but for the moment it's easy to
add new hash to relcache.
Fixed small bug in xlog.c:ReadRecord.
as well allow DROP multiple INDEX, RULE, TYPE as well. Add missing
CommandCounterIncrement to DROP loop, which could cause trouble otherwise
with multiple DROP of items affecting same catalog entries. Try to
bring a little consistency to various error messages using 'does not exist',
'nonexistent', etc --- I standardized on 'does not exist' since that's
what the vast majority of the existing uses seem to be.
* Makefile: Add more standard targets. Improve shell redirection in GNU
make detection.
* src/backend/access/transam/rmgr.c: Fix incorrect(?) C.
* src/backend/libpq/pqcomm.c (StreamConnection): Work around accept() bug.
* src/include/port/unixware.h: ...with help from here.
* src/backend/nodes/print.c (plannode_type): Remove some "break"s after
"return"s.
* src/backend/tcop/dest.c (DestToFunction): ditto.
* src/backend/nodes/readfuncs.c: Add proper prototypes.
* src/backend/utils/adt/numutils.c (pg_atoi): Cope specially with strtol()
setting EINVAL. This saves us from creating an extra set of regression test
output for the affected systems.
* src/include/storage/s_lock.h (tas): Correct prototype.
* src/interfaces/libpq/fe-connect.c (parseServiceInfo): Don't use variable
as dimension in array definition.
* src/makefiles/Makefile.unixware: Add support for GCC.
* src/template/unixware: same here
* src/test/regress/expected/abstime-solaris-1947.out: Adjust whitespace.
* src/test/regress/expected/horology-solaris-1947.out: Part of this file
was evidently missing.
* src/test/regress/pg_regress.sh: Fix shell. mkdir -p returns non-zero if
the directory exists.
* src/test/regress/resultmap: Add entries for Unixware.
and DropBuffers. Formerly we cleared the flag for each buffer currently
belonging to the target rel or database, but that's completely wrong!
Must look at BufferTagLastDirtied to see whether the BufferDirtiedByMe
flag is relevant to target rel or not; this is *independent* of the
current contents of the buffer. Vadim spotted this problem, but his
fix was only partially correct...
> Regression tests opr_sanity and sanity_check are now failing.
Um, Bruce, I've said several times that I didn't think Perchine's large
object changes should be applied until someone had actually reviewed
them.
Makefile.port, since they are of no use to configure and much of the
library magic happens in Makefile.port anyway.
Use __alpha, not __alpha__, since the former is universally available.
Remove -DNOFIXADE from the compile command line and put it in the port
include file.
I tested it restoring my database with > 100000 BLOBS, and dumping it out.
But unfortunatly I can not restore it back due to problems in pg_dump.
--
Sincerely Yours,
Denis Perchine
source directory. This involves mostly makefiles using $(srcdir) when they
might have used ".". (Regression tests don't work with this, yet.)
Sort out usage of CPPFLAGS, CFLAGS (and CXXFLAGS). Add "override" keyword
in most places, to preserve necessary flags even when the user overrode the
flags.
anywhere else that Makefile.shlib needs to modify CFLAGS to produce
valid code for a shared library. I'm not real clear on *why* the use
of override causes make to ignore the later attempt to assign
CFLAGS +=
but it indubitably does --- at least on gmake 3.79.1. gmake bug?
This patch forces the use of 'DROP VIEW' to destroy views.
It also changes the syntax of DROP VIEW to
DROP VIEW v1, v2, ...
to match the syntax of DROP TABLE.
Some error messages were changed so this patch also includes changes to the
appropriate expected/*.out files.
Doc changes for 'DROP TABLE" and 'DROP VIEW' are included.
--
Mark Hollomon
* doc/src/sgml/installation.sgml: ditto.
* src/test/regress/README: Regenerate.
* doc/src/sgml/docguide.sgml: Explain how it was done. Explain how
INSTALL and HISTORY are (now) generated.
* doc/src/sgml/Makefile: Implement HISTORY generation to be analoguous
to INSTALL.
I prepared and tested a patch vs. 7.0.2, and it works fine. I've added
another option which allows users to have their own service file in
~/.pg_service.conf, which might come handy sometimes.
Mario Weilguni
I have noticed that there are some identical files in
postgresql-7.0.2/src/test/regress/expected/
> diff float8-cygwin.out float8-small-is-zero.out #I recommend deleting
float8-cygwin.out
> diff geometry-cygwin-precision.out geometry-solaris-precision.out #I
recommend deleting geometry-cygwin-precision.out
below is the diff of postgresql-7.0.2/src/test/regress/resultmap
that has the above files deleted plus the addition of an alpha regression
test built with alphaev56-dec-osf4.0e/2.95.2/ . The alpha geometry
regression file is attached
11c11
< float8/i.86-pc-cygwin*=float8-cygwin
---
> float8/i.86-pc-cygwin*=float8-small-is-zero
18c18
< geometry/i.86-pc-cygwin*=geometry-cygwin-precision
---
> geometry/i.86-pc-cygwin*=geometry-solaris-precision
21a22
> geometry/alpha.*-dec-osf=geometry-alpha-precision
Ricardo Muggli
Systems Manager
Information and Technology Services
Minnesota State University, Mankato
Only two have shown up on the web site. Even the mbox is missing the
second.
The missing patch is a one-liner, so here it is. I can resend the
whole bug report if wanted.
Pete Forman
As a result, backend/libpq/pqcomm.c and interfaces/libpq/fe-connect.c
fail to compile.
The <netinet/tcp.h> header needs to be preceded by <netinet/in.h>, at
least on IRIX, Solaris and AIX. The simple configure test fails.
(That header on Linux is idempotent.)
The basic problem is that <netinet/tcp.h> is a BSD header. The
correct header for TCP internals such as TCP_NODELAY on a UNIX system
is <xti.h>. By UNIX I mean UNIX95 (aka XPG4v2 or SUSv1) or later.
The current UNIX standard (UNIX98 aka SUSv2) is available online at
<http://www.opengroup.org/onlinepubs/7908799/>.
The fix is to add header support for <xti.h> into configure.in and
config.h.in.
The 2 files which conditionally include <netinet/tcp.h> need also to
conditionally include <xti.h>.
Pete Forman
I have counted 9 differences in the least significant digit compared
with geometry-positive-zeros.out. It would be wise for someone else
to double check.
Pete Forman
it previously. The patch included is against fairly current sources, but
it may apply cleanly against 7.0.2 as well.
On Fri, 6 Oct 2000, Vilson farias wrote:
> I found a irregular behavior with constraints.
>
> I can only set a referencial integrity between these tables when there are
> no data, even if there are no change to referential integrity violation.
couldn't produce a full patch using cvs diff -c this time since I have
created new files and anonymous cvs usage doesn't allow you to
adds. I'm supplying the modified src/interfaces/jdbc as a tarball at :
http://www.candleweb.no/~gunnar/projects/pgsql/postgres-jdbc-2000-10-05.tgz
The new files that should be added are :
? org/postgresql/PGStatement.java
? org/postgresql/ObjectPool.java
? org/postgresql/ObjectPoolFactory.java
There is now a global static pool of free byte arrays and used byte arrays
connected to a statement object. This is the role of the new PGStatement
class. Access to the global free array is synchronized, while we rely on
the PG_Stream synchronization for the used array.
My measurements show that the perfomance boost on this code is not quite as
big as my last shot, but it is still an improvement. Maybe some of the
difference is due to the new synchronization on the global array. I think I
will look into choosing between on a connection level and global level.
I have also started experimented with improving the performance of the
various conversions. The problem here is ofcourse related handle the
various encodings. One thing I found to speed up ResultSet.getInt() a lot
was to do custom conversion on the byte array into int instead of going
through the getString() to do the conversion. But I'm unsure if this is
portable, can we assume that a digit never can be represented by more than
one byte ? It works fine in my iso-latin-8859-1 environment, but what about
other environments ? Maybe we could provide different ResultSet
implementations depending on the encoding used or delegate some methods of
the result set to an "converter class".
Check the org/postgresql/jdbc2/FastResultSet.java in the tarball above to
see the modified getInt() method.
Regards,
Gunnar
for the library. not sure if this will cause problems on other
platforms, but if it does it can be easily fixed. Also remove the
references to the GeekGadgets includes as the majority of users don't
have them installed and they foul the build process. We can document
that adding them if you have them installed is a good idea.
David Reid
> > For a while I though it might be because we are using an alpha TAS in
> > the spinlock rather than the old semaphore. I replaced our spinlock
> > with the standard one and it made no difference. We have been running
> > with our spinlock implementation for nearly 2 months on a production
> > database now without a hitch, so I think it is ok. Did I ever submit
> > any patches for the Alpha spinlock?
>
> Not that I recall. We did get some advice from some Alpha gurus at DEC
> who seemed to think the existing TAS code is OK. What was it that you
> felt needed to be improved?
The current code uses semaphores, which has the advantage that it works
well even on multi-processor machines, but the disadvantage that it is not
the fastest way possible. Writing a spinlock on Alpha for SMP machines is
very difficult, as you need to deal with memory barriers. A real mess. But
then one of the people at Compaq pointed out to us that there is a
ready-made routine on Alpha. We implemented it with the two patches below.
I ran tests with lots of parallel back-ends and got around a 10% speed
increase. I include the two patches. Perhaps some of the other people
running Tru64 can have a look at these as well.
Cheers,
Adriaan Joubert
> this is patch v 0.4 to support transactions with BLOBs.
> All BLOBs are in one table. You need to make initdb.
>
> --
> Sincerely Yours,
> Denis Perchine
after that dynamic loading isn't working and shared memory handling is
broken.
Attached with this message, there is a Zip file which contain :
* beos.diff = patch file generated with difforig
* beos = folder with beos support files which need to be moved in /
src/backend/port
* expected = foler with three file for message and precision
difference in regression test
* regression.diff = rule problem (need to kill the backend manualy)
* dynloader = dynloader files (they are also in the pacth files,
but there is so much modification that I have join full files)
Everything works except a problem in 'rules' Is there some problems
with rules in the current tree ? It used to works with last week tree.
Cyril VELTER
took some rejiggering of typename and ACL parsing, as well as moving
parse_analyze call out of parser(). Restructure postgres.c processing
so that parse analysis and rewrite are skipped when in abort-transaction
state. Only COMMIT and ABORT statements will be processed beyond the raw
parser() phase. This addresses problem of parser failing with database access
errors while in aborted state (see pghackers discussions around 7/28/00).
Also fix some bugs with COMMIT/ABORT statements appearing in the middle of
a single query input string.
Function, operator, and aggregate arguments/results can now use full
TypeName production, in particular foo[] for array types.
DROP OPERATOR and COMMENT ON OPERATOR were broken for unary operators.
Allow CREATE AGGREGATE to accept unquoted numeric constants for initcond.
SQL92 semantics, including support for ALL option. All three can be used
in subqueries and views. DISTINCT and ORDER BY work now in views, too.
This rewrite fixes many problems with cross-datatype UNIONs and INSERT/SELECT
where the SELECT yields different datatypes than the INSERT needs. I did
that by making UNION subqueries and SELECT in INSERT be treated like
subselects-in-FROM, thereby allowing an extra level of targetlist where the
datatype conversions can be inserted safely.
INITDB NEEDED!
working on the VERY latest version of BeOS. I'm sure there will be
alot of comments, but then if there weren't I'd be disappointed!
Thanks for your continuing efforts to get this into your tree.
Haven't bothered with the new files as they haven't changed.
BTW Peter, the compiler is "broken" about the bool define and so on.
I'm filing a bug report to try and get it addressed. Hopefully then we
can tidy up the code a bit.
I await the replies with interest :)
David Reid
problems with some bits of it, but when all the patches are in it'll build
and we can fix it from there :) I've got a version that builds and runs and
that is the basis for these patches.
The first file has the new additional files that are required,
template/beos
backend/port/dynloader/beos.c
backend/port/dynloader/beos.h
include/port/beos.h
makefiles/Makefile.beos
The second is a tarball of diffs against a few files. I've added sys/ipc.h
to configure and config.h via configure.in and config.h.in and then started
adding the check as this file isn't needed on BeOS and having loads of
#ifdef BEOS isn't as obvious as #ifdef HAVE_SYS_IPC_H and isn't as
autconf'ish :)
Files touched are
include/c.h
configure.in
include/config.h.in
include/storage/ipc.h
include/utils/int8.h
Let me know how these go. I'll await a response before submitting any more.
Any problems just get in touch.
David Reid
have already modified "next_insert()" in 7.0-ecpglib. However
in the meaning of speed-up, the patch will be needed.
--
Regards,
SAKAIDA Masaaki -- Osaka, Japan
would close and then re-open rel being truncated. Depending on the
luck of the draw, the re-opened relcache entry might or might not be
at the same physical location as before. Unfortunately, if it wasn't
then heap_truncate would crash and burn, because it still had a pointer
at the old location. Fix is to open and then close rel in
RelationTruncateIndexes, so that rel's refcount never goes to zero
until heap_truncate is done.
and the fmgr redesign.
It makes the homebrewn dl*() functions for more recent Versions of AIX
obsolete
by using the system dl*() functions instead.
It also fixes the expected file for the horology regression test.
Please regenerate configure from configure.in, I don't have the
environment/time.
Andreas
(Don't forget that an alias is required.) Views reimplemented as expanding
to subselect-in-FROM. Grouping, aggregates, DISTINCT in views actually
work now (he says optimistically). No UNION support in subselects/views
yet, but I have some ideas about that. Rule-related permissions checking
moved out of rewriter and into executor.
INITDB REQUIRED!
Update the installation instructions (formerly misnamed "FAQ"), add configure
checks for some headers rather than having users copy stubs manually (ugh!).
Use Autoconf check for exe extension. This also avoids inheriting the value
of $(X) from the environment.
illegal call to pg_mbclipen() that is for backend only. However I
have not remove the entire part of the problem, rather mark it with
#ifdef MULTIBYTE_NOTUSED since we should come back with a long range
solution someday.
add --without-tk option to disable Tk. We don't need the AC_PATH_XTRA
test because tkConfig.sh already contains all the information about how to
compile and link with X. Also make sure that libpq is up to date for
libpgtcl. Remove executable bits from pgaccess.sh, but add it to pgaccess.
complaints about ungrouped variables. This is for consistency with
behavior elsewhere, notably the fact that the relname is reported as
an alias in these same complaints. Also, it'll work with subselect-
in-FROM where old code didn't.
- rename ichar() to chr() (discussed with Tom)
- add docs for oracle compatible routines:
btrim()
ascii()
chr()
repeat()
- fix bug with timezone in to_char()
- all to_char() variants return NULL instead textin("")
if it's needful.
The contrib/odbc is without changes and contains same routines as main
tree ... because I not sure how plans are Thomas with this :-)
Karel
---------------------------------------------------------------------------
This effectively one line patch should fix the fact that
foreign key definitions in create table were erroring if
a primary key was defined. I was using the columns
list to get the columns of the table for comparison, but
it got reused as a temporary list inside the primary key
stuff.
Stephan Szabo
query representation. Note that GEQO_RELS setting is now interpreted
as the number of top-level items in the FROM list, not necessarily the
number of relations in the query. This seems appropriate since we are
only doing join-path searching over the top-level items.
There is still no effective difference but it will kick in once setuid
functions exist (not included here). Make old getpgusername() alias for
current_user.
ie, consider only the columns coming from the JOIN clause's sub-clauses.
Also detect attempts to reference columns belonging to other tables
(which would still be possible using an explicitly-qualified name).
I'm not sure this implements the spec's semantics 100% accurately, but
at least it gives plausible behavior.
DESTDIR=/else/where' and prepends the value of DESTDIR to the full
installation paths (e.g., /else/where/usr/local/pgsql/bin). This allows
users to install the package into a location different from the one that
was configured and hard-coded into various scripts, e.g., for creating
binary packages.
DESTDIR is in many cases preferrable over `make install
prefix=/else/where' because
a) `prefix' affects the path that is hard-coded into the files, which can
lead to a `make install prefix=xxx' (as done by the regression test
driver) corrupting the files in the source tree with wrong paths.
b) it doesn't work at all if a directory was overridden to not depend on
`prefix', e.g., --sysconfdir=/etc.
(Updating the regression test driver to use DESTDIR is a separate
undertaking.)
See also autoconf@gnu.org, From: Akim Demaille <akim@epita.fr>, Date: 08
Sep 2000 12:48:59 +0200, Message-ID:
<mv4em2vb1lw.fsf@nostromo.lrde.epita.fr>, Subject: Re: HTML format
documentation.
- Use symbols for tests on relkind (ie. use RELKIND_VIEW, not 'v')
- Fix bug in support for -b option (== --blobs).
- Dump views as views (using 'create view').
- Remove 'isViewRule' since we check the relkind when getting tables.
- Now uses temp table 'pgdump_oid' rather than 'pg_dump_oid' (errors otherwise).
- Added extra param for specifying handling of OID=0 and which typename to output.
- Fixed bug in SQL scanner when SQL contained braces. (in rules)
- Use format_type function wherever possible
regression tests for specific types, and move a few others to the
cross-type "horology" test.
Rearrange the test order slightly, and move the abstime test to the
"parallel safe" area.
Hand-patch the results for "1947" and for "solaris", so those may not
be exactly correct.
one-line change necessary. Due to the Mark Holloman "New Relkind for
Views" patch, my support for views in the driver will need to be updated
to match. The change to DatabaseMetaData.getTableTypes[][] is as
follows:
- {"VIEW", "(relkind='r' and relhasrules='t' and relname !~
'^pg_' and relname !~ '^xinv')"},
+ {"VIEW", "(relkind='v' and relname !~ '^pg_' and relname
!~ '^xinv')"},
Christopher Cain
driver if the translations files have not been properly installed. (We
carefully avoided installing the translations file in a controlled
environment here specifically to test for such a bug. :-)
See attached description for more details.
William
--
William Webber william@peopleweb.net.au
This is a patch which lets the DatabaseMetaData return the object type
when getTables(....) is called. It does not really fix any bug, but it
fills in some functionality that should be there anyway. The diff
included here is off of the CVS as of just now :)
----------------------------------------------------------------
Travis Bauer | CS Grad Student | IU |www.cs.indiana.edu/~trbauer
----------------------------------------------------------------
for views. Views are now have a "relkind" of
RELKIND_VIEW instead of RELKIND_RELATION.
Also, views no longer have actual heap storage
files.
The following changes were made
1. CREATE VIEW sets the new relkind
2. The executor complains if a DELETE or
INSERT references a view.
3. DROP RULE complains if an attempt is made
to delete a view SELECT rule.
4. CREATE RULE "_RETmytable" AS ON SELECT TO mytable DO INSTEAD ...
1. checks to make sure mytable is empty.
2. sets the relkind to RELKIND_VIEW.
3. deletes the heap storage files.
5. LOCK myview is not allowed. :)
6. the regression test type_sanity was changed to
account for the new relkind value.
7. CREATE INDEX ON myview ... is not allowed.
8. VACUUM myview is not allowed.
VACUUM automatically skips views when do the entire
database.
9. TRUNCATE myview is not allowed.
THINGS LEFT TO THINK ABOUT
o pg_views
o pg_dump
o pgsql (\d \dv)
o Do we really want to be able to inherit from views?
o Is 'DROP TABLE myview' OK?
--
Mark Hollomon
7.0.2 release. Sorry, if that's fixed ages ago - I don't track
development versions of PostgreSQL.
Patch is just a little bit tested (some valid functions created and
successfully run as well as some erroneous ones created and emitted proper
error messages when used).
My platform is FreeBSD 5.0-CURRENT (with perl 5.6.0 provided in the
base system).
Alex Kapranoff
encountered while getting my reporting tool up and running with the
driver. All changes are in the DatabaseMetaData class.
Problem: The getDatabaseProductVersion() method was returning "6.5.2"
Resolution: Changed it to return "7.0.2"
Problem: A call to getTables() with an unsupported table type (in the
String array) resulted in a malformed SQL statement and subsequent
parsing error
Resolution: Unsupported table types are now ignored without error
Problem: In a getTables() call, tables and views were both returned by
the "TABLE" table type, and the "VIEW" table type was unsupported
Resolution: Changed the "TABLE" type to return only physical tables and
added support for the "VIEW" table type (returning only views)
Problem: The getIdentifierQuoteString() method was returning null
Resolution: This method now returns a double-quote
Christopher Cain
Here's the multibyte aware version of my patch to fix the truncation
of the rulename autogenerated during a CREATE VIEW. I've modified all
the places in the backend that want to construct the rulename to use
the MakeRetrieveViewRuleName(), where I put the #ifdef MULTIBYTE, so
that's the only place that knows how to construct a view rulename. Except
pg_dump, where I replicated the code, since it's a standalone binary.
The only effect the enduser will see is that views with names len(name)
> NAMEDATALEN-4 will fail to be created, if the derived rulename clases
with an existing rule: i.e. the user is trying to create two views with
long names whose first difference is past NAMEDATALEN-4 (but before
NAMEDATALEN: that'll error out after the viewname truncation.) In no
case will the user get left with a table without a view rule, as the
current code does.
Ross Reedstrom
The regression tests abstime, horology, int2, int4, and tinterval fail
by default. They will pass if comparison is made to one of the
variant expected/*.out files.
(A related problem is that the geometry test fails. This seems to be
just a matter of FP precision but more rigorous inspection is needed.
No existing expected/geometry-*.out variant matches.)
Pete Forman
user is now defined in terms of the user id, the user name is only computed
upon request (for display purposes). This is kind of the opposite of the
previous state, which would maintain the user name and compute the user id
for permission checks.
Besides perhaps saving a few cycles (integer vs string), this now creates a
single point of attack for changing the user id during a connection, for
purposes of "setuid" functions, etc.
quote_ident(text) returns text
quote_literal(text) returns text
These are handy to build up properly quoted query strings
for the new PL/pgSQL EXECUTE functionality to submit
dynamic DDL statements.
Jan
=# create table t (id int4 unique);
NOTICE: CREATE TABLE/UNIQUE will create implicit index 't_id_key' for table 't'
=# begin;
query: drop table t;
NOTICE: Caution: DROP TABLE cannot be rolled back, so don't abort now
NOTICE: Caution: DROP INDEX cannot be rolled back, so don't abort now
=# rollback;
=# drop table t;
NOTICE: mdopen: couldn't open t: No such file or directory
NOTICE: RelationIdBuildRelation: smgropen(t): No such file or directory
NOTICE: mdopen: couldn't open t: No such file or directory
NOTICE: mdopen: couldn't open t: No such file or directory
NOTICE: mdopen: couldn't open t_id_key: No such file or directory
NOTICE: RelationIdBuildRelation: smgropen(t_id_key): No such file or directory
NOTICE: mdopen: couldn't open t: No such file or directory
NOTICE: RelationIdBuildRelation: smgropen(t): No such file or directory
NOTICE: mdopen: couldn't open t: No such file or directory
ERROR: cannot open relation t
- full support for IW (ISO week) and vice versa conversion for IW too
(the to_char 'week' support is now complete and I hope correct).
Thomas, I use for IW code from timestamp.c, for this I create separate
function date2isoweek() from original 'case DTK_WEEK:' code in the
timestamp_part(). I mean will better use one code for same feature in
date_part() and in to_char(). The isoweek2date() is added to timestamp.c
too. Right?
IMHO in 7.1 will all to_char's features complete. It is cca 41 templates
for date/time and cca 21 for numbers.
* to_ascii:
- gcc, is it correct now? :-)
In the patch is documentation for to_char's IW and for to_ascii().
Karel
didn't hear anything about, but which would
have broken with the function manager changes
anyway.
Well, this patch checks that a unique constraint
of some form (unique or pk) is on the referenced
columns of an FK constraint and that the columns
in the referencing table exist at creation time.
The former is to move closer to SQL compatibility
and the latter is in answer to a bug report.
I also added a basic check of this functionality
to the alter table and foreign key regression
tests.
Stephan Szabo
sszabo@bigpanda.com
incarnations (I hope). When an acceptable flex version is not found, print
instructive error messages from both configure and the makefiles, so that
users can continue building anyway.
virtual FDs, we just return the ENFILE/EMFILE error to the caller,
rather than immediate elog(). This allows more robust behavior in
the postmaster, which uses AllocateFile() but does not want elog().
length is < TOAST_TUPLE_THRESHOLD, even with toastable column types
present. For example, CREATE TABLE foo (f1 int, f2 varchar(100))
does not require a toast table, even though varchar is a toastable
type.
pg_mb2wchar(const unsigned char *, pg_wchar *);
pg_mb2wchar_with_len(const unsigned char *, pg_wchar *, int);
from void to int. Now they return the number of
wide chars.
(rather than compile time). For libpq, even when Kerberos support is
compiled in, the default user name should still fall back to geteuid()
if it can't be determined via the Kerberos system.
A couple of fixes for string type configuration parameters, now that there
is one.
for example, an SQL function can be used in a functional index. (I make
no promises about speed, but it'll work ;-).) Clean up and simplify
handling of functions returning sets.
right circumstances a hash join executed as a DECLARE CURSOR/FETCH
query would crash the backend. Problem as seen in current sources was
that the hash tables were stored in a context that was a child of
TransactionCommandContext, which got zapped at completion of the FETCH
command --- but cursor cleanup executed at COMMIT expected the tables
to still be valid. I haven't chased down the details as seen in 7.0.*
but I'm sure it's the same general problem.
pg_proc.c (where it's actually used). Fix it to correctly handle tlists
that contain resjunk target items, and improve error messages. This
addresses bug reported by Krupnikov 6-July-00.
macros where appropriate (the code used to have several different ways
of doing that, including Int32, Int8, UInt8, ...). Remove last few
references to float32 and float64 typedefs --- it's all float4/float8
now. The typedefs themselves should probably stay in c.h for a release
or two, though, to avoid breaking user-written C functions.
Update functions to new-style fmgr, make BIT and VARBIT be binary-
equivalent, add entries to allow these types to be btree indexed,
correct a few bugs. BIT/VARBIT are now toastable, too.
NOTE: initdb forced due to catalog updates.
right thing with variable-free clauses that contain noncachable functions,
such as 'WHERE random() < 0.5' --- these are evaluated once per
potential output tuple. Expressions that contain only Params are
now candidates to be indexscan quals --- for example, 'var = ($1 + 1)'
can now be indexed. Cope with RelabelType nodes atop potential indexscan
variables --- this oversight prevents 7.0.* from recognizing some
potentially indexscanable situations.
including utility statements. Still can't copy or compare executor
state, but at present that doesn't seem to be necessary. This makes
it possible to execute most (all?) utility statements in plpgsql.
Had to change parsetree representation of CreateTrigStmt so that it
contained only legal Nodes, and not bare string constants.
from Param nodes, per discussion a few days ago on pghackers. Add new
expression node type FieldSelect that implements the functionality where
it's actually needed. Clean up some other unused fields in Func nodes
as well.
NOTE: initdb forced due to change in stored expression trees for rules.
as MaxHeapAttributeNumber. Increase MaxAttrSize to something more
reasonable (given what it's used for, namely checking char(n) declarations,
I didn't make it the full 1G that it could theoretically be --- 10Mb
seemed a more reasonable number). Improve calculation of MaxTupleSize.
that RAND_MAX applies to them, since it doesn't. Instead add a
config.h parameter MAX_RANDOM_VALUE. This is currently set at 2^31-1
but could be auto-configured if that ever proves necessary. Also fix
some outright bugs like calling srand() where srandom() is appropriate.
rather than the "~~" operator; this made it easy to add ESCAPE features.
Implement ILIKE, NOT ILIKE, and the ESCAPE clause for them.
afaict this is not MultiByte clean, but lots of other stuff isn't either.
Fix up underlying support code for LIKE/NOT LIKE.
Things should be faster and does not require internal string copying.
Update regression test to add explicit checks for
LIKE/NOT LIKE/ILIKE/NOT ILIKE.
Remove colon and semi-colon operators as threatened in 7.0.
Implement SQL99 COMMIT/AND NO CHAIN.
Throw elog(ERROR) on COMMIT/AND CHAIN per spec
since we don't yet support it.
Implement SQL99 CREATE/DROP SCHEMA as equivalent to CREATE DATABASE.
This is only a stopgap or demo since schemas will have another
implementation soon.
Remove a few unused production rules to get rid of warnings
which crept in on the last commit.
Fix up tabbing in some places by removing embedded spaces.
rather than the "~~" operator; this made it easy to add ESCAPE features.
Implement ILIKE, NOT ILIKE, and the ESCAPE clause for them.
afaict this is not MultiByte clean, but lots of other stuff isn't either.
Fix up underlying support code for LIKE/NOT LIKE.
Things should be faster and does not require internal string copying.
Update regression test to add explicit checks for
LIKE/NOT LIKE/ILIKE/NOT ILIKE.
Remove colon and semi-colon operators as threatened in 7.0.
Implement SQL99 COMMIT/AND NO CHAIN.
Throw elog(ERROR) on COMMIT/AND CHAIN per spec
since we don't yet support it.
Implement SQL99 CREATE/DROP SCHEMA as equivalent to CREATE DATABASE.
This is only a stopgap or demo since schemas will have another
implementation soon.
Remove a few unused production rules to get rid of warnings
which crept in on the last commit.
Fix up tabbing in some places by removing embedded spaces.
that giving pg_proc a toast table required solving the same problems
we'd have to solve for pg_class --- pg_proc is one of the relations
that gets bootstrapped in relcache.c. Solution is to go back at the
end of initialization and read in the *real* pg_class row to replace
the phony entry created by formrdesc(). This should work as long as
there's no need to touch any toasted values during initialization,
which seems a reasonable assumption.
Although I did not add a toast-table for every single system table
with a varlena attribute, I believe that it would work to just do
ALTER TABLE pg_class CREATE TOAST TABLE. So anyone who's really
intent on having several thousand ACL entries for a rel could do it.
NOTE: I didn't force initdb, but you must do one to see the effects
of this patch.
thing when there are multiple result relations. Formerly, during
something like 'UPDATE foo*', foo's constraints and *only* foo's
constraints would be applied to all foo's children. Wrong-o ...
multiple times in the parsetree (can happen in COALESCE or BETWEEN
contexts, for example). This is a pretty grotty solution --- it will
do for now, but perhaps we can do better when we redesign querytrees.
What we need is a consistent policy about whether querytrees should be
considered read-only structures or not ...
- encode 'text' from database encoding to ASCII
to_ascii('\256\341k')
to_ascii( text, int4 )
- encode 'text' from 'int4' encoding to ASCII
to_ascii('\256\341k', 8)
to_ascii( text, name )
- encode 'text' from 'name' encoding to ASCII
to_ascii('\256\341k', 'LATIN2')
Now is supported LATIN1, LATIN2, WIN1250. For other character sets I
haven't good resources. Add new encoding is easy...
If encoding is not supported returns ERROR.
Note --- not exists total corect conversion to ASCII, this function try
convert chars those is _probably_ interpret-able in ASCII for
others use ' '. But for example for all Czech characters it is
sufficient ... hmm Chinese / JAP and other complicated langs
have
bad luck here :-(
Karel
trying to toast tuples inserted into toast tables! Fix is two-pronged:
first, ensure all columns of a toast table are marked attstorage='p',
and second, alter the target chunk size so that it's less than the
threshold for trying to toast a tuple. (Code tried to do that but the
expression was wrong.) A few cosmetic cleanups in tuptoaster too.
NOTE: initdb forced due to change in toaster chunk-size.
on myself to do something about the non-self-consistency of the inet
comparison functions. The results are probably still semantically wrong
(inet and cidr should have different comparison semantics, I think)
but at least the boolean operators now agree with each other and with
the sort order of indexes on inet/cidr.
These two routines will now ALWAYS elog() on failure, whether you ask for
a lock or not. If you really want to get a NULL return on failure, call
the new routines heap_open_nofail()/heap_openr_nofail(). By my count there
are only about three places that actually want that behavior. There were
rather more than three places that were missing the check they needed to
make under the old convention :-(.
result, in fact nearly the opposite of what it should, because it
was passing the not-equal operator to eqsel() which would use it to
compare the value against the most common value in the column, and
of course obtain the wrong result therefrom. Must pass the equality
operator to eqsel() instead. Fortunately that's easy to get from
the oprnegate link.
At this point I think it'd be possible to make float4 be pass-by-value
without too much work --- and float8 too on machines where Datum is
8 bytes. Something to try when the mood strikes, anyway.
- Added code to dump 'Create Schema' statement (pg_dump)
- Don't bother to disable/enable triggers if we don't have a superuser (pg_restore)
- Cleaned up code for reconnecting to database.
- Force a reconnect as superuser before enabling/disabling triggers.
- Added & Removed --throttle (pg_dump)
- Fixed minor bug in language dumping code: expbuffres were not being reset.
- Fixed version number initialization in _allocAH (pg_backup_archiver.c)
- Added second connection when restoring BLOBs to allow temp. table to survive
(db reconnection causes temp tables to be lost).
(Sorry, couldn't help it...)
Removed type filename as well, since it's unused and probably useless.
INITDB FORCED, because pg_rewrite columns are now plain text again.
allows fixing problems with operators that expected to be able to
return a NULL, such as the '#' line-segment-intersection operator
that tried to return NULL when the two segments don't intersect.
(See, eg, bug report from 1-Nov-99 on pghackers.) Fix some other
bugs in passing, such as backwards comparison in path_distance().
I did not force. I marked numeric as compressable-but-not-move-off-able,
partly to test that storage mode and partly because I've got doubts
that numerics are large enough to need external storage.
Note that this has changed some of the edge cases for what is accepted
as a type name and/or column id. Regression test passes, but more
tweaks may be coming...
the planner may try to generate them as a result of transitivity of the
existing int2-vs-int4 and int4-vs-int8 operators. In fact, it is now
necessary that mergejoinable cross-datatype operators form closed sets.
Add an opr_sanity regress test to detect missing operators.
FreeBSD/Intel and DecUX/Alpha machines. The bug appears in postgresql
6.5.3 and 7.0.2. Can someone please review it and apply it to the
source tree?
Sometimes when the postgres connection dies it is necessary to
attempt to reconnect. Calling the pgconnection::Connect method in a
derived class leaks memory because it does not clear the current
connection (if there is one). These patches ensures that any open
connections are closed before attempting to open a new one.
-Michael Richards
to use with a multiple-key index. Formerly we would only extract clauses
that had to do with the first key of the index, which was correct but
didn't exploit the index fully.
actually, but who could understand it with no comments? Fix bug
while at it: _bt_orderkeys would try to invoke comparisons on
NULL inputs, given the right sort of redundant quals.
mergejoinable qual clauses, and add them to the query quals. For
example, WHERE a = b AND b = c will cause us to add AND a = c.
This is necessary to ensure that it's safe to use these variables
as interchangeable sort keys, which is something 7.0 knows how to do.
Should provide a useful improvement in planning ability, too.
varlena elements work now. Allow assignment to previously-nonexistent
subscript position to extend array, but only for 1-D arrays and only
if adjacent to existing positions (could do more if we had a way to
represent nulls in arrays, but I don't want to tackle that now).
Arrange for assignment of NULL to an array element in UPDATE to be a
no-op, rather than setting the entire array to NULL as it used to.
(Throwing an error would be a reasonable alternative, but it's never
done that...) Update regress test accordingly.
work as expected. THe underlying implementation is essentially
'SET foo = array_set(foo, 1, bar)', so we have to turn the items
into nested invocations of array_set() to make it work correctly.
Side effect: we now complain about 'UPDATE tab SET foo = bar, foo = baz'
which is illegal per SQL92 but we didn't detect it before.
Remove a bunch of crufty code for large-object-based arrays, which is
superseded by TOAST and likely hasn't worked in a long time anyway.
Clean up array code a little, and in particular eliminate its habit
of scribbling on the input array (ie, modifying the input tuple :-().
left keys during bottom-up index build, and leave some free space
instead of packing the pages to the brim (so as to avoid vast numbers
of page splits during the first interactive insertions).
- Support for BLOB output from pg_dump and input via pg_restore
- Support for direct DB connection in pg_restore
- Fixes in support for --insert flag
- pg_dump now outputs in modified OID order
- Support for direct DB connection in pg_restore
- Fixes in support for --insert flag
- pg_dump now outputs in modified OID order
- various other bug fixes
duplicate keys by letting search go to the left rather than right when an
equal key is seen at an upper tree level. Fix poor choice of page split
point (leading to insertion failures) that was forced by chaining logic.
Don't store leftmost key in non-leaf pages, since it's not necessary.
Don't create root page until something is first stored in the index, so an
unused index is now 8K not 16K. (Doesn't seem to be as easy to get rid of
the metadata page, unfortunately.) Massive cleanup of unreadable code,
fix poor, obsolete, and just plain wrong documentation and comments.
See src/backend/access/nbtree/README for the gory details.
The latter updated accordingly. Also add `dist' and `distcheck' targets
to play with, but caveat packager.
Updated backend/bootstrap and backend/parser makefile to make them
marginally builddir aware and fix the usual set of things.
Add rule to automatically remake config.h dependent on config.h.in and
config.status. (Adopted from Autoconf manual and about every other
package.) On a good day we should now have a complete and accurate set
of dependencies throughout everything.
in a non-safe interpreter, so with full OS access! Language is
restricted to be used by DB superusers.
Added "argisnull n" and "return_null" commands to gain full control
over NULL values from new FMGR capabilities.
Jan
type different from input type but are expecting ExecAgg to insert the
first non-null input as the starting transition value. This has always
been verboten, but wasn't checked for until now...
NOTE: this implementation of tcl_avg() fails with 'divide by zero'
for zero input rows. It ought to return NULL, but pltcl does not
currently provide a way to do that, so I'm leaving the problem unsolved
for now.
documentation. Therefore it's now installed by default. If there is no
documentation to be found (i.e., you are not using the distribution)
then this step is skipped.
Add --docdir option to configure to control installation directory.
There's now only one transition value and transition function.
NULL handling in aggregates is a lot cleaner. Also, use Numeric
accumulators instead of integer accumulators for sum/avg on integer
datatypes --- this avoids overflow at the cost of being a little slower.
Implement VARIANCE() and STDDEV() aggregates in the standard backend.
Also, enable new LIKE selectivity estimators by default. Unrelated
change, but as long as I had to force initdb anyway...
AlterTableAddConstraint. The major changes from the last patch
are that it should hopefully check for references to temp tables
(not in the shadow case, but at defination time) from permanent tables in
foreign keys and refuse them and that it doesn't allow the table(s)
being constrained to be views (because those cases don't currently
work).
Stephan SzaboThis should be a slighly more complete patch for commands/command.c
AlterTableAddConstraint. The major changes from the last patch
are that it should hopefully check for references to temp tables
(not in the shadow case, but at defination time) from permanent tables in
foreign keys and refuse them and that it doesn't allow the table(s)
being constrained to be views (because those cases don't currently
work).
Stephan Szabo
pass-by-ref data types --- eg, an index on lower(textfield) --- no longer
leak memory during index creation or update. Clean up a lot of redundant
code ... did you know that copy, vacuum, truncate, reindex, extend index,
and bootstrap each basically duplicated the main executor's logic for
extracting information about an index and preparing index entries?
Functional indexes should be a little faster now too, due to removal
of repeated function lookups.
CREATE INDEX 'opt_type' clause is deimplemented by these changes,
but I haven't removed it from the parser yet (need to merge with
Thomas' latest change set first).
Include updates for the comment.sql regression test.
Implement SET SESSION CHARACTERISTICS and SET DefaultXactIsoLevel.
Implement SET SESSION CHARACTERISTICS TRANSACTION COMMIT
and SET AutoCommit in the parser only.
Need to add code to actually do something.
Implement WITHOUT TIME ZONE type qualifier.
Define SCHEMA keyword, along with stubbed-out grammar.
Implement "[IN|INOUT|OUT] [varname] type" function arguments
in parser only; INOUT and OUT throws an elog(ERROR).
Add PATH as a type-specific token, since PATH is in SQL99
to support schema resource search and resolution.
them, but forgot to attach relevant restriction clauses, so that the
plan represented a scan over the whole table with restrictions applied
as qpquals not indexquals. Another day, another bug...
* the result is not recorded anywhere
* the result is not used anywhere
* the result is only used in some places, whereas others have been getting away with it
* the result is used improperly
Also make command line options handling a little better (e.g., --disable-locale,
while redundant, should really still *dis*able).
memory contexts. Currently, only leaks in expressions executed as
quals or projections are handled. Clean up some old dead cruft in
executor while at it --- unused fields in state nodes, that sort of thing.
in-chunk leaks, overwrite-next-chunk leaks and overwrite block-freeptr leaks.
A in-chunk leak --- if something overwrite space after wanted (via palloc()
size, but it is still inside chunk. For example
x = palloc(12); /* create 16b chunk */
memset(x, '#', 13);
this leak is in the current source total invisible, because chunk is 16b and
leak is in the "align space".
For this feature I add data_size to StandardChunk, and all memory which go
from AllocSetAlloc() is marked as 0x7F.
The MemoryContextCheck() is compiled '#ifdef USE_ASSERT_CHECKING'.
I add this checking to 'tcop/postgres.c' and is active after each backend
query, but it is probably not sufficient, because some MemoryContext exist
only during memory processing --- will good if someone who known where
it is needful (Tom:-) add it for others contexts;
A problem in the current source is that we have still some malloc()
allocation that is not needful and this allocation is total invisible for
all context routines. For example Dllist in backend (pretty dirty it is in
catcache where values in Dllist are palloc-ed, but list is malloc-ed).
--- and BTW. this Dllist design stand in the way for query cache :-)
Tom, if you agree I start replace some mallocs.
BTW. --- Tom, have you idea for across transaction presistent allocation for
SQL functions? (like regex - now it is via malloc)
I almost forget. I add one if() to AllocSetAlloc(), for 'size' that are
greater than ALLOC_BIGCHUNK_LIMIT is not needful check AllocSetFreeIndex(),
because 'fidx' is always 'ALLOCSET_NUM_FREELISTS - 1'. It a little brisk up
allocation for very large chunks. Right?
Karel
* Add option to build with OpenSSL out of the box. Fix thusly exposed
bit rot. Although it compiles now, getting this to do something
useful is left as an exercise.
* Fix Kerberos options to defer checking for required libraries until
all the other libraries are checked for.
* Change default odbcinst.ini and krb5.srvtab path to PREFIX/etc.
* Install work around for Autoconf's install-sh relative path anomaly.
Get rid of old INSTL_*_OPTS variables, now that we don't need them
anymore.
* Use `gunzip -c' instead of g?zcat. Reportedly broke on AIX.
* Look for only one of readline.h or readline/readline.h, not both.
* Make check for PS_STRINGS cacheable. Don't test for the header files
separately.
* Disable fcntl(F_SETLK) test on Linux.
* Substitute the standard GCC warnings set into CFLAGS in configure,
don't add it on in Makefile.global.
* Sweep through contrib tree to teach makefiles standard semantics.
... and in completely unrelated news:
* Make postmaster.opts arbitrary options-aware. I still think we need to
save the environment as well.
backend functions via backend PQexec(). The SPI interface has long
been our only documented way to do this, and the backend pqexec/portal
code is unused and suffering bit-rot. I'm putting it out of its misery.
Does not work since it fetches one byte beyond the source data, and when
the phase of the moon is wrong, the source data is smack up against the
end of backend memory and you get SIGSEGV. Don't laugh, this is a fix
for an actual user bug report.
|> developers so we are sure it will work on all platforms.
The problem with the current settings is that the linker is called
directly. This is wrong, it should always be called through the
compiler
driver (the only exception is `ld -r'). This will make sure that the
necessary libraries like libgcc are linked in.
But there is still a different problem with the setting of LDFLAGS_ODBC.
The psqlodbc module defines the functions _init and _fini which are
reserved for the shared library initialisation. These should be changed
to constructor functions. Then LDFLAGS_ODBC can be changed to be just
`-lm'. Btw, why does it use -Bsymbolic?
Andreas Schwab
functional.
Handle include file installation in src/include/Makefile
genbki.sh improvements: Don't substitute anything by config.status,
instead pass in AWK and CPP through environment. Change calling
convention to support named output files, so we get to see error
messages on stderr.
Rename bootstrap template files and install them into PREFIX/share.
Update initdb to that effect and other readability improvements
in initdb.
- The problems Jan reported
- incompatibility with configure (now uses HAVE_LIBZ instead of HAVE_ZLIB)
- a problem in auto-detecting archive file format on piped archives
Philip Warner
Special handling of TOAST relations during VACUUM. TOAST relations
are vacuumed while the lock on the master table is still active.
The ANALYZE flag doesn't propagate to their vacuuming because the
toaster access routines allways use index access ignoring stats, so
why compute them at all.
Protection of TOAST relations against normal INSERT/UPDATE/DELETE
while offering SELECT for debugging purposes.
Jan
PostgreSQL-7.0.2 run on Linux for the Intel-IA64 architecture. It also
fixes a bug in the configure scripts that caused configure to fail on
the fcntl(F_SETLK) test.
This fix triggered a bug in the fcntl(F_SETLK) code of the Linux
kernel when used on unix domain sockets resulting in postmaster to
segfault immediately after startup. There is a fix available and
included in the kernel that will be on SuSE Linux 7.0, but kernels <=
2.2.16 still have this bug.
Reinhard Max
files to restrict the set of users that can connect to a database
but can still use the pg_shadow password. (You just leave off the
password field in the secondary file.)
Don't go through pg_exec_query_dest(), but directly to the execution
routines. Also, extend parameter lists so that there's no need to
change the global setting of allowSystemTableMods, a hack that was
certain to cause trouble in the event of any error.
COPYs of pg_shadow and pg_group.
It also turns out that pg_dumpall was all but broken for multiple servers
running at non-standard port numbers. You might get the users and groups
from one server and the databases from another. Fixed that.
A little user interface and code cleanup along with that. This also takes
care of the portability bug discussed in "[BUGS] pg_dumpall" in March 2000.
and config.h. Adjusted all referring code.
Scrapped pg_version and changed initdb accordingly. Integrated
src/utils/version.c into src/backend/utils/init/miscinit.c. Changed all
callers.
Set version number to `7.1devel'. (Non-numeric version suffixes now allowed.)
Now the to_timestamp() support WW,W,J,SSSS,DDD conversion from strings and
the am/pm bug is fixed, the to_char() use week-of-year (WW) full compatible
with Oracle.
This patch update relevant regress-tests and docs too.
Karel
~
~
Don't use DISABLE_COMPLEX_MACRO on Solaris. Don't define the
replacement function in the header file. Use -KPIC, not -K PIC.
Use CC to link C++ libraries, not ld/ar.
Eliminate file not found warnings in tcl build code.
entry that has rules. This allows us to release the rule parsetrees
on relcache flush without needing a working freeObject() routine.
Formerly, the rule trees were leaked permanently at relcache flush.
Also, clean up handling of rule creation and deletion --- there was
not sufficient locking of the relation being modified, and there was
no reliable notification of other backends that a relcache reload
was needed. Also, clean up relcache.c code so that scans of system
tables needed to load a relcache entry are done in the caller's
memory context, not in CacheMemoryContext. This prevents any
un-pfreed memory from those scans from becoming a permanent memory
leak.
in copyfuncs and equalfuncs exposed by regression tests. We still have
some work to do: these modules really ought to handle most or all of
the utility statement node types. But it's better than it was.
documentation. Let's try to keep this file a bit neater in future,
hmm? Also (to get back to the original point) update info about
FUNC_MAX_ARGS, and add additional config symbols for debugging
new memory management changes.
worth the effort to continue to maintain. Since freeObject() is not
capable of coping with cases like multiple links to a node, it's
unlikely that it ever will be useful again. We now have memory
context management that offers a faster and more reliable way of
getting rid of arbitrary node trees (at the cost of having to know
in advance of building the tree that you'll want to get rid of it).
standard targets and behaviour. Replaced Makefile.in's with
Makefile's and declared the respective variables in Makefile.global.
maintainer-clean target now available at top level, although it does
not work in the backend tree yet.
Cleanup pass over Makefile.shlib, renamed some targets and variables.
The shared library symlink tests are now done by make, not the shell.
ecpg: Remove one warning in sloppy flex output.
PL/Perl and Perl interface: the MakeMaker documentation is confusing,
the realclean target *does* "delete derived files", but it also
uninstalls them. Don't use that.
The submake targets in the various bin directories that update libpq
should `make all', not `make libpq.a'. That is a) unportable, and
b) doesn't build the shared library.
for details). It doesn't really do that much yet, since there are no
short-term memory contexts in the executor, but the infrastructure is
in place and long-term contexts are handled reasonably. A few long-
standing bugs have been fixed, such as 'VACUUM; anything' in a single
query string crashing. Also, out-of-memory is now considered a
recoverable ERROR, not FATAL.
Eliminate a large amount of crufty, now-dead code in and around
memory management.
Fix problem with holding off SIGTRAP, SIGSEGV, etc in postmaster and
backend startup.
option settings. Sort out SIGHUP vs BACKEND -- there is no total ordering
here, so make explicit checks. Add comments explaining all of this.
Removed permissions check on SHOW command.
Add examine_subclass to the game, rename to SQL_inheritance to fit the
official data model better. Adjust documentation.
Standalone backend needs to reset all options before it starts. To
facilitate that, have IsUnderPostmaster be set by the postmaster itself,
don't wait for the magic -p switch.
Also make sure that all environment variables and argv's survive
init_ps_display(). Use strdup where necessary.
Have initdb make configuration files (postgresql.conf, pg_hba.conf) mode
0600 -- having configuration files is no fun if you can't edit them.
to apply the tempname->realname mapping to type name lookup as well
as relation name lookup, else the type tuple will not be found when
wanted. This fixes bugs like this one:
create temp table foo (f1 int);
select foo.f2 from foo;
ERROR: Unable to locate type name 'foo' in catalog
are opened in a consistent order by different backends (I ordered them
by index OID because that's easy, but any other consistent order would
do as well). This avoids potential deadlock for index types that we
acquire exclusive locks on ... ie, rtree.
entries now for int8 and network hash indexes. int24_ops and int42_ops
are gone. pg_opclass no longer contains multiple entries claiming to be
the default opclass for the same datatype. opr_sanity regress test
extended to catch errors like these in the future.
materialized tupleset is small enough) instead of a temporary relation.
This was something I was thinking of doing anyway for performance, and Jan
says he needs it for TOAST because he doesn't want to cope with toasting
noname relations. With this change, the 'noname table' support in heap.c
is dead code, and I have accordingly removed it. Also clean up 'noname'
plan handling in planner --- nonames are either sort or materialize plans,
and it seems less confusing to handle them separately under those names.
passing the index-is-unique flag to index build routines (duh! ...
why wasn't it done this way to begin with?). Aside from eliminating
an eyesore, this should save a few milliseconds in btree index creation
because a full scan of pg_index is not needed any more.
discussion of 5/19/00). pg_index is now searched for indexes of a
relation using an indexscan. Moreover, this is done once and cached
in the relcache entry for the relation, in the form of a list of OIDs
for the indexes. This list is used by the parser and executor to drive
lookups in the pg_index syscache when they want to know the properties
of the indexes. Net result: index information will be fully cached
for repetitive operations such as inserts.
pointers, namely the catcache tuple fetch routines. Also get rid of
the unused and possibly confusing 'size' field in struct cachedesc.
Since it doesn't allow for variable-length fields, anyone who
actually trusted it would likely be making a mistake...
was inappropriately relying on rel->rd_nblocks to tell if the LO is
empty (apparently a hack to get around a long-dead index bug), causing
misbehavior on a written-but-never-vacuumed LO. Also, inv_read failed
to cope gracefully with 'holes' (unwritten regions) in the object.
we'll get there one day.
Use `cat' to create aclocal.m4, not `aclocal'. Some people don't
have automake installed.
Only run the autoconf rule in the top-level GNUmakefile if the
invoker specified `make configure', don't run it automatically
because of CVS timestamp skew.
--- ie, they're only called for side-effects. Add a PG_RETURN_VOID()
macro and use it where appropriate. This probably doesn't change the
machine code by a single bit ... it's just for documentation.
> situation is already tracked in File routines, but a little bit
> incorrectly.
> After small survey in Linux kernel code, I am not sure about
> it. New patch set pos to unknown in the case of read/write
> fails. And do lseek again.
> Here is the full patch for this. This patch reduce amount of
> lseek call ten ti mes for update statement and twenty times for
> select statement. I tested joined up date and count(*) select
> for table with rows > 170000 and 10 indices. I think this is
> worse of trying. Before lseek calls account for more than 5% o
> f time. Now they are 0.89 and 0.15 respectevly.
>
> Due to only one file modification patch should be applied in
> src/backedn/stora ge/file/ dir.
-- Sincerely Yours,
Denis Perchine
have'nt r un autoconf to create a new configure, I guess that's done by
the smapshot process, I had to remove a line from interface/odbc/
GNUMakefile to get it to build, it was a autoconf variable that looks to
not be used anymore, I am assuming that this is ok.
Nick Gorham
Easysoft Ltd
>> Makefile where the make bombs if "." is not in the builder's path?
>> The last I checked, it wasn't applied and the fix is very easy
>> (explicitly use "./" to call the script).
SL Baur
quote-stripping, and acl-checking tasks for these functions from the
parser, and do them at function execution time instead. This fixes
the failure of pg_dump to produce correct output for nextval(Foo)
used in a rule, and also eliminates the restriction that the argument
of these functions must be a parse-time constant.
Interfaced a lot of the custom tests to the config.cache, in the process
made them separate macros and grouped them out into files. Made naming
adjustments.
Removed a couple of useless/unused configure tests.
Disabled C++ by default. C++ is no more special than Perl, Python, and Tcl.
And it breaks equally often. :(
that now functions as a wrapper around the MakeMaker stuff. It might
even behave sensically when we have separate build dirs. Same for plperl,
which of course still doesn't work very well. Made sure that plperl
respects the choice of --libdir.
Added --with-python to automatically build and install the Python interface.
Works similarly to the Perl5 stuff.
Moved the burden of the distclean targets lower down into the source tree.
Eventually, each make file should have its own.
Added automatic remaking of makefiles and configure. Currently only for the
top-level because of a bug(?) in Autoconf. Use GNU `missing' to work around
missing autoconf and aclocal. Start factoring out macros into their own
config/*.m4 files to increase readability and organization.
absolute. It also makes it more compliant with the interface
specification in Sun's documentation;
1. absolute(0) should throw an exception.
2. absolute(>num-records) should set the current row to after the last
record in addition to returning false.
3. absolute(<num-records) should set the current row to before the first
record in addition to returning false.
These operations in the existing code just return false and don't change
current_row.
These changes required a minor change to relative(int) since it calls
absolute(int)
The attached patch is against the cvs repository tree as of this morning.
Also, who is in charge of maintaining the jdbc driver? I'm working on
getArray for the jdbc2 driver, but it's going to require three more
classes to be added to the driver, and thus three more source files
in the repository. Is there someone I can contact directly to ask about
this?
Travis Bauer | CS Grad Student | IU |www.cs.indiana.edu/~trbauer
postgres build and use unixODBC (http://www.unixodbc.org)
This patch was applied against the postgresql-7.0beta1 build
Any problems let me know.
Nick Gorham
more restriction for fretful users. The current PG allow define only
NO-CREATE-DB and NO-CREATE-USER restriction, but for some users I need
NO-CREATE-TABLE and NO-LOCK-TABLE.
This patch add to current code NOCREATETABLE and NOLOCKTABLE feature:
CREATE USER username
[ WITH
[ SYSID uid ]
[ PASSWORD 'password' ] ]
[ CREATEDB | NOCREATEDB ] [ CREATEUSER | NOCREATEUSER ]
-> [ CREATETABLE | NOCREATETABLE ] [ LOCKTABLE | NOLOCKTABLE ]
...etc.
If CREATETABLE or LOCKTABLE is not specific in CREATE USER command,
as default is set CREATETABLE or LOCKTABLE (true).
A user with NOCREATETABLE restriction can't call CREATE TABLE or
SELECT INTO commands, only create temp table is allow for him.
Karel
to_char. I don't know about the rest of the world, but the "standard" in
Australia is the following:
1st, 2nd, 3rd, 4th - 9th
10th - 19th
21st, 22nd, 23rd, 24th - 29th (similarly for 30s - 90s)
110th - 119th (and for all "teens")
121st, 122nd, 123rd, 124th - 129th
I think you see the trend. The current code works fine except that it
produces:
111st, 112nd, 113rd, 114th - 119th
211st, 212nd, 213rd, 214th - 219th ... and so on.
Without knowing anything about what's supported (and what isn't) in the usual
I18N libraries, should this type of behaviour be defined within the locales?
Daniel Baldoni
It addresses three issues:
1. The problem with ResultSet's interface specifying 1-based indexing was
not quite fixed in 7.0.2. absolute would stop the user form moving to the
first record (record 0 internally).
2. Absolute did not set current_row
3. For field.mod=-1, GetObject would try to return numeric values with a
precision of around 65000. Now GetObject detects when field.mod==-1, and
passes that as the scale to getBigDecimal. getBigDecimal detects when a
-1 is passed and simply does not scale the value returned. You still get
the correct value back, it simply does not tweak the precision.
I'm working off of a source tree I just checked out from the
repository. The diff is based on what was in the repository about ten
minutes ago.
----------------------------------------------------------------
Travis Bauer | CS Grad Student | IU |www.cs.indiana.edu/~trbauer
----------------------------------------------------------------
CPP) to create platform independent files. Unfortunately, that means that
every config.status (or configure) run invariably causes a relink of the
postmaster and also that we can't put these files in the distribution
(usefully). So we make it a little smarter: when the output files already
exist and it notices that it would recreate them in identical form, it
doesn't touch them. In order to avoid re-running the make rule all the time
we update a timestamp file instead.
Update release_prep accordingly. Also make Gen_fmgrtab.sh use the awk that
is detected at configure time, not necessarily named `awk' and have it check
for exit statuses a little better.
In other news... Remove USE_LOCALE from the templates, it was set to `no'
everywhere anyway. Also remove YACC and YFLAGS from the templates, configure
is smart enough to find bison or yacc itself. Use AC_PROG_YACC for that
instead of the hand-crafted code. Do not set YFLAGS to `-d'. The make rules
that need this flag should explicitly invoke it. YFLAGS should be a user
variable. Update the makefiles to that effect.
over multiple lookups --- it should use SearchSysCacheTupleCopy instead.
This accounts for rare failures like 'init_fcache: null probin for procedure 481'
when running concurrently with a VACUUM.
direct pointer into the syscache entry for the type. In some cases
the syscache entry might get flushed before we are done using the
returned type name. This bug accounts for difficult-to-repeat
failures seen when INSERTs into columns of certain data types are
run in parallel with VACUUMs of system tables. There may be related
problems elsewhere --- we need to take a harder look at uses of
syscache data.
Fixed Statement, so that the update count is valid when an SQL DELETE operation is done.
While fixing the update count, made it easier to get the OID of the last insert as well. Example is in example/basic.java
inputs have been converted to newstyle. This should go a long way towards
fixing our portability problems with platforms where char and short
parameters are passed differently from int-width parameters. Still
more to do for the Alpha port however.
no reason for them to be copied into src/backend rather than being
installed straight from the catalog subdirectory. This also avoids
some peculiar behavior (bugs?) present in at least gmake 3.78.1: it
won't always update the bki files in backend/ even when the ones in
backend/catalog/ are newer.
that name and issue a NOTICE to the effect that we did. Previously,
code would try to assign the new cursor declaration to the old portal,
but this didn't work reliably since new parsetree is still sitting in
blank portal and is likely to get clobbered.
actually use their targetlist, are given a targetlist that is just a
pointer to the first appended plan's targetlist. This is OK, but what
is not OK is that any sub-select expressions in said tlist were being
entered in the subPlan lists of both the Append and the first appended
plan. That led to two startup and two shutdown calls for the same
plan node at exec time, which led to crashes. Fix is to not generate
a list of subPlans for an Append node. Same problem and fix apply
to other node types that don't have a real, functioning targetlist:
Material, Sort, Unique, Hash.
it will close VFDs if necessary to surmount ENFILE or EMFILE failures.
Make use of this in md.c, xlog.c, and user.c routines that were
formerly vulnerable to these failures. In particular, this should
handle failures of mdblindwrt() that have been observed under heavy
load conditions. (By golly, every other process on the system may
crash after Postgres eats up all the kernel FDs, but Postgres will
keep going!)
(ie, parameters instead of consts) will be treated as a range query.
We do not know the actual selectivities involved, but it seems like
a good idea to use a smaller estimate than we would use for two unrelated
inequalities.
That means you can now set your options in either or all of $PGDATA/configuration,
some postmaster option (--enable-fsync=off), or set a SET command. The list of
options is in backend/utils/misc/guc.c, documentation will be written post haste.
pg_options is gone, so is that pq_geqo config file. Also removed were backend -K,
-Q, and -T options (no longer applicable, although -d0 does the same as -Q).
Added to configure an --enable-syslog option.
changed all callers from TPRINTF to elog(DEBUG)
built-in procedures are named after the prosrc field of pg_proc (ie,
the actual C function name), not the proname field. This did not use
to make a difference back when the two were always the same, but in the
presence of overloaded proname values we'd best try to use the C name
instead. AFAICT this change affects no existing code, but it is
necessary to be able to get at some built-in functions that no macro
was being generated for before.
to 10, and be consistent about whether it counts the trailing null (it
does not). Also increase MAXDATELEN to be sure no buffer overflows are
caused by the longer MAXTZLEN.
key call sites are changed, but most called functions are still oldstyle.
An exception is that the PL managers are updated (so, for example, NULL
handling now behaves as expected in plperl and plpgsql functions).
NOTE initdb is forced due to added column in pg_proc.
Most (nearly all) of the work was done by David Wragg <dpw@doc.ic.ac.uk>
He patched 6.5.3. I've updated it for 7.0RC5.
It works for MIT kerberos 1.1.1 (and previously for 1.0.6 as well).
I've got the patch against 6.5.3, plus kerberized RPMS.
Mike Wyer <mw@doc.ic.ac.uk> || "Woof?"
other than the most common value in a column. We had had 0.5, make it
0.1 to make it more likely that an indexscan will be chosen. Really
need better statistics instead, but this should stem the bleeding
meanwhile ...
subsequent I/O attempts fail cleanly. I'm speculating about failure
scenarios in which we do pq_close, then something in a proc_exit routine
opens a file (re-using that kernel FD number), then something else
fails and tries to write an elog message to the frontend ... message
ends up in opened file, oops. No known examples of this but it seems
like a potential hole.
the oper field should be a valid Node structure so it can be dumped by
outfuncs.c without risk of coredump. (We had been using a raw pointer
to character string, which surely is NOT a valid Node.) This doesn't
cause any backwards compatibility problems for stored rules, since
raw unanalyzed parsetrees are never stored.
*last*, after all updating of system catalogs. In old code, an error
detected during TypeRename left the relation hosed. Also, add a call
to flush the relation's relcache entry, rather than trusting to shared
cache invalidation to flush it for us.
think that both sides of indexqual look like index keys. An example is
create table inside (f1 float8 primary key);
create table outside (g1 float8, g2 float8);
select * from inside,outside where f1 = atan2(g1+1, g2);
ERROR: ExecInitIndexScan: both left and right ops are rel-vars
(note that failure is potentially platform-dependent). Solution is a
cleanup I had had in mind to make anyway: functional index keys should
be represented as Var nodes in the fixed indexqual, just like regular
index keys.
it exists) before testing 'using namespace std'. This is necessary
on some C++ setups where the compiler won't take a 'using' until
you've included a header that mentions namespace std. (Pretty braindead
if you ask me, but...)
--with-includes) to makefiles for pltcl and plperl, so that these
switches will be used even though we do not want other top-level
CFLAGS. Ain't it fun trying to support multiple-compiler platforms?
project I am working on (Recall - a distributed, fault-tolerant,
replicated, storage framework @ http://www.fault-tolerant.org).
Recall is written in C++. I need to include the postgres headers and
there are some problems when including the headers w/C++.
Attached is a patch generated from postgres/src that fixes my problems.
I was hoping to get this into the main source. It's very small (2k) and
3 files are changed: backend/utils/fmgr/fmgr.c,
backend/utils/Gen_fmgrtab.sh.in, and include/access/tupdesc.h.
In C++, you get a multiply defined symbol because the variable
(FmgrInfo *fmgr_pl_finfo) is defined in the header (the patch moves it
to the .c file). The other problem in tupdesc.h is the use of typeid
is a problem in c++ (I renamed it to oidtypeid).
Thanks,
Neal Norwitz
some platforms --- and I also see that it is documented as not thread-
safe on HPUX and possibly other platforms. No good reason not to just
use IPPROTO_TCP constant from <netinet/in.h> instead.
really ought to fix relcache entry construction so that it does not
do so much with CurrentMemoryContext = CacheCxt. As is, relatively
harmless leaks in either sequential or index scanning translate to
permanent leaks if they occur when called from relcache build.
For the moment, however, the path of least resistance is to repair
all such leaks...
Hiroshi. ReleaseRelationBuffers now removes rel's buffers from pool,
instead of merely marking them nondirty. The old code would leave valid
buffers for a deleted relation, which didn't cause any known problems
but can't possibly be a good idea. There were several places which called
ReleaseRelationBuffers *and* FlushRelationBuffers, which is now
unnecessary; but there were others that did not. FlushRelationBuffers
no longer emits a warning notice if it finds dirty buffers to flush,
because with the current bufmgr behavior that's not an unexpected
condition. Also, FlushRelationBuffers will flush out all dirty buffers
for the relation regardless of block number. This ensures that
pg_upgrade's expectations are met about tuple on-row status bits being
up-to-date on disk. Lastly, tweak BufTableDelete() to clear the
buffer's tag so that no one can mistake it for being a still-valid
buffer for the page it once held. Formerly, the buffer would not be
found by buffer hashtable searches after BufTableDelete(), but it would
still be thought to belong to its old relation by the routines that
sequentially scan the shared-buffer array. Again I know of no bugs
caused by that, but it still can't be a good idea.
RowExclusive (my fault). Also, install a check to prevent people
from trying COPY BINARY to stdout/from stdin. No way that will
work unless we redesign the frontend COPY protocol ... which is
not worth the trouble in the near future ...
is in <string> and not in <string.h> on QNX4/egcs-2.91.60.
Probably this can be changed for all platforms. The test in line 1705 uses
<string> as well. Because I am not sure, I havn't this included into the
patch.
doc/Makefile has to be sligthly modified as it has been done for
src/backend/Makefile due to a QNX4 problem (patch attached)
Furthermore src/test/regress/run_check.sh needs to be patched as it has been
done for regress.sh (patch attached). Please note that in the patch the
postmaster is started always with the -i option.
run_check.sh reports the test "limit" as failed, but in reallity it is OK.
regress.sh reports it as OK.
Andreas Kardos
IRIX systems using the native compilers. A summary is:
- Various files use "//" as a comment delimiter in c files.
- Problems caused by assuming "char" is signed.
cash.in: building -signed the rules regression test fails as described
in FAQ_QNX4. If CHAR_MAX is "255U" then ((signed char)CHAR_MAX) is -1.
postmaster.c: random number regression test failed without this change.
- Some generic build issues and warning message cleanup.
David Kaelbling
just use the portable form,
tr ABCDEFGHIJKLMNOPQRSTUVWXYZ abcdefghijklmnopqrstuvwxyz
There were a bunch of places that weren't paying attention to configure's
result anyway (including configure itself!?); clean them up too.
I'm including a diff of
postgresql-7.0/src/interfaces/jdbc/org/postgresql/jdbc2/ResultSet.java.
I've clearly marked all the fixes I did. Would *someone* who has access
to the cvs please put this in?
Joseph Shraibman
days. It seems to be a FAQ, and I think I know why. When creating a 'c'
language function, CREATE FUNCTION is fed the shared object filename,
and seems to succeed. Only when trying to use the function is an error
thrown, by which time the coder thinks something's wrong with executing
the code, not with loading it.
I think I once saw it proposed to load shared objects at function creation
time, but that idea was shot down on the grounds of resident memory bloat,
ISTR. Here's a patch for a compromise: all it does is stat() the file,
just like the loader code does, so that the errors caused by non existent
files, and no directory 'x' permissions (the most common ones, it seems),
get caught while the developer is still thinking about code loading. It
doesn't catch all errors (like the code not being readable by the postgres
user) but seems to catch the most common, without actually opening the file.
What do you think?
Ross
indexes, apparently, nor on functional indexes with more than one input
column (force of natts = 1 was in the wrong branch of IF statement).
Coredumped if source relation contained any uncommitted tuples, due to
failure to test for success return from heap_fetch. Fetched tuple
was passed directly to heap_insert, which clobbers the TID and commit
status in the tuple header it's given, which meant that the source
relation's tuples all got trashed as the copy proceeded. Abort partway
through, and you're left with a lot of missing tuples.
I wonder what else is lurking here ...
under FreeBSD ... basically, if setproctitle() exists, use it ...
the draw back right now is the PS_SET_STATUS stuff doesn't work, but am looking
into that one right now ... at lesat now you can see who is connecting where
and from where ...
mappings. In fact, it had them backward because it was using the 6.5.*
code. Copied them from parser/gram.y, so it is fixed now. Looks like
our first 7.0.1 fix. Oops, seems Tom has beat me to it as I was typing
this.
Rearrange handling of VACUUMs so that they are certain to be executed
as superuser not some random user; also, do not forget to vacuum
template1 itself.
pg_char_to_encoding() in multibyte disbaled case so that it does not
throw an error, rather return HARD CODED default value (currently SQL_ASCII).
This would solve the "non-mb backend vs. mb-enabled frontend" problem.
cleanup, ie, as soon as we have caught the longjmp. This ensures that
current context will be a valid context throughout error cleanup. Before
it was possible that current context was pointing at a context that would
get deleted during cleanup, leaving any subsequent pallocs in deep
trouble. I was able to provoke an Assert failure when compiled with
asserts + -DCLOBBER_FREED_MEMORY, if I did something that would cause
an error to be reported by the backend large-object code, because indeed
that code operates in a context that gets deleted partway through xact
abort --- and CurrentMemoryContext was still pointing at it! Boo hiss.
cases where joinclauses were present but some joins have to be made
by cartesian-product join anyway. An example is
SELECT * FROM a,b,c WHERE (a.f1 + b.f2 + c.f3) = 0;
Even though all the rels have joinclauses, we must join two of them
in cartesian style before we can use the join clause...
than BIND_DEFERRED. That way, if the loaded library has unresolved
references, shl_load fails cleanly. As we had it, shl_load would
succeed and then the dynlinker would call abort() when we try to call
into the loaded library. abort()ing a backend is uncool.
always failed if Perl makefile's INSTALLSITELIB variable was specified
in terms of another variable. Fix by adding an echo-installdir target
to the Perl makefile, which the upper-level Makefile can invoke.
libpq++.h contained copies of the class declarations in the other libpq++
include files, which was bogus enough, but the declarations were not
completely in step with the real declarations. Remove these in favor
of including the headers with #include. Make PgConnection destructor
virtual (not absolutely necessary, but seems like a real good idea
considering the number of subclasses derived from it). Give all classes
declared private copy constructors and assignment operators, to prevent
compiler from thinking it can copy these objects safely.
compiler than the one selected to build Postgres with. It was trying
to feed Postgres-compiler switches to Tcl's compiler. (Seen this before
with the perl5 interface...) Fix to use only CFLAGS taken from Tcl's
configure information, plus -I which is pretty universal.
unless you feed it -Aa or -Ae switch. Autoconf does not know about this,
but we can fix it in the hpux_cc template file. I knew templates were
good for something ;-)
to give wrong results: it should be looking at inJoinSet not inFromCl.
Also, make 'modified' flag be local to ApplyRetrieveRule: we should
append a rule's quals to the query iff that particular rule applies,
not if we have fired any previously-considered rule for the query!
(SELECT FROM table*). Cause was reference to 'eref' field of an RTE,
which is null in an RTE loaded from a stored rule parsetree. There
wasn't any good reason to be touching the refname anyway...
table for an average of NTUP_PER_BUCKET tuples/bucket, but cost_hashjoin
was assuming a target load of one tuple/bucket. This was causing a
noticeable underestimate of hashjoin costs.
(LIKE and regexp matches). These are not yet referenced in pg_operator,
so by default the system will continue to use eqsel/neqsel.
Also, tweak convert_to_scalar() logic so that common prefixes of strings
are stripped off, allowing better accuracy when all strings in a table
share a common prefix.
subsequent elogs() in the same COPY operation to display the wrong
line number. Fix is to clear lineno only when elog level is such
that we will not return to caller.
repaired psql option scanning bug (special treatment to \g |pipe)
fixed ipcclean makefile
made configure look for Perl to handle psql help build gracefully
extern int inet_aton(const char *cp, struct in_addr * addr);
appearing before the optional #define for const, which was certain
to fail on a machine with neither const nor inet_aton().
compiler will understand them. configure may have #define'd them to
empty because the local C compiler doesn't understand them, but this
may very well cause a C++ compilation to fail, so don't do it in C++.
all platforms, not just SCO. The operation is undefined for Unix-domain
sockets anyway. It seems SCO is not the only platform that complains
instead of treating the call as a no-op.
include the version from backend/port into libpq.
There is a second-rate implementation of inet_aton() already present
in fe-connect.c, #ifdef'd WIN32. That ought to be removed in favor
of using the better version from port/. However, since I'm not in a
position to test the WIN32 code, I will leave well enough alone for
this release...
contained a sub-SELECT nested within an AND/OR tree that cnfify()
thought it should rearrange. Same physical sub-SELECT node could
end up linked into multiple places in resulting expression tree.
This is harmless for most node types, but not for SubLink.
Repair bug by making physical copies of subexpressions that get
logically duplicated by cnfify(). Also, tweak the heuristic that
decides whether it's a good idea to do cnfify() --- we don't really
want that to happen when it would cause multiple copies of a subselect
to be generated, I think.
whether to do fsync or not, and if so (which should be seldom) just
do the fsync immediately. This way we need not build data structures
in md.c/fd.c for blind writes.
logged queries to 1024, truncating longer queries. That is about half of
the size I need (I have a union that is 2K long). Can someone consider
bumping it to 4K or so? Patch attached...
Regards,
Ed Loehr
as a shared dirtybit for each shared buffer. The shared dirtybit still
controls writing the buffer, but the local bit controls whether we need
to fsync the buffer's file. This arrangement fixes a bug that allowed
some required fsyncs to be missed, and should improve performance as well.
For more info see my post of same date on pghackers.
than not knowing what they are at all. Perhaps they should have their own
type category? Hard to say. In the meantime, doing it this way allows
SELECT 'unknown' || 'unknown' to continue being resolved as textcat,
instead of spitting out an ambiguous-operator error.
parse node types. This allows these statements to be placed in a plpgsql
function. Also, see to it that statement types not handled by the copy
logic will draw an appropriate elog(ERROR), instead of leaving a null
pointer that will cause coredump later on. More utility statements could
be added if anyone felt like turning the crank.
Add a random number generator and seed setter (random(), SET SEED)
Fix up the interval*float8 math to carry partial months
into the time field.
Add float8*interval so we have symmetry in the available math.
Fix the parser and define.c to accept SQL92 types as field arguments.
Fix the parser to accept SQL92 types for CREATE TYPE, etc. This is
necessary to allow...
Bit/varbit support in contrib/bit cleaned up to compile and load
cleanly. Still needs some work before final release.
Implement the "SOME" keyword as a synonym for "ANY" per SQL92.
Implement ascii(text), ichar(int4), repeat(text,int4) to help
support the ODBC driver.
Enable the TRUNCATE() function mapping in the ODBC driver.
Ensure that outer tuple link needed for inner indexscan qual evaluation
gets set in the EvalPlanQual case. This stops coredump, but we still
have resource leaks due to failure to clean up EvalPlanQual properly...
It does work with the following patch applied and gcc 2.95.2 .
Use --with-template=aix_gcc to compile the whole lot with gcc.
The geometry regression test produces different precision.
With optimization I run into regression failures starting at oidjoins,
thus no -O2. Anybody else try gcc 2.95.2 and -O2 on beta4 ?
This is an important patch, since recent versions of the IBM compiler
are not for free, and thus most questions I get concern gcc.
Andreas
PS.: I am testing with beta4
xact abort state in pg_exec_query_dest, we should continue scanning the
querytree list, on the off chance that one of the later queries in the
string is COMMIT or ROLLBACK.
would crash, due to premature invocation of SetQuerySnapshot(). Clean
up problems with handling of multiple queries by splitting
pg_parse_and_plan into two routines. The old code would not, for
example, do the right thing with END; SELECT... submitted in one query
string when it had been in transaction abort state, because it'd decide
to skip planning the SELECT before it had executed the END. New
arrangement is simpler and doesn't force caller to plan if only
parse+rewrite is needed.
be an expression not just a simple Var, so long as only one table is
referenced (so that code isn't really any more difficult than before).
This whole thing is still fundamentally bogus, but at least we can accept
a few more cases than before.
WHERE in a place where it can be part of a nestloop inner indexqual.
As the code stood, it put the same physical sub-Plan node into both
indxqual and indxqualorig of the IndexScan plan node. That confused
later processing in the optimizer (which expected that tracing the
subPlan list would visit each subplan node exactly once), and would
probably have blown up in the executor if the planner hadn't choked first.
Fix by making the 'fixed' indexqual be a complete deep copy of the
original indexqual, rather than trying to share nodes below the topmost
operator node. This had further ramifications though, because we were
making the aforesaid list of sub-Plan nodes during SS_process_sublinks
which is run before construction of the 'fixed' indexqual, meaning that
the copy of the sub-Plan didn't show up in that list. Fix by rearranging
logic so that the sub-Plan list is built by the final set_plan_references
pass, not in SS_process_sublinks. This may sound like a mess, but it's
actually a good deal cleaner now than it was before, because we are no
longer dependent on the assumption that planning will never make a copy
of a sub-Plan node.
Should be more robust to overflows.
Pass through an unmapped function unchanged, rather than rejecting it.
Add a few more functions, but comment out those which can go through as-is.
Can be used with contrib/odbc/ package, though that isn't committed yet.
here is an updated version of the bit type with a bugfix and all the necessa
ry
SQL functions defined. This should replace what is currently in contrib. I'd
appreciate any comments on what is there.
Kind regards,
Adriaan
pg_internal.init file in-place, which meant that if another backend
started at about the same time, it might read the incomplete file.
init_irels tries to guard against that, but I have now seen a crash
due to reading bad data from a partly-written file. (This may indicate
a kernel bug on my platform? Not sure.) Anyway, clearly the safest
course is to write the new pg_internal.init file under a unique temporary
filename, and rename it into place only after it's all written.
- I was unable to compile ecpg due to the ":=" instead of "=" in defining
LIBPQDIR and some other variables in Makefile.global.in
- pg_id (and also pg_encoding) executable was not removed during "make
clean" - there was no $(X) appended to the executable name for rm
- I have added result for int2, int4, float8 and geometry regression tests
- int2, int2 - yet another message for too large numbers ;-)
- float8 - it is problably a bug in the newlib C library - it has no
error message for numbers with exponent -400
- geometry - differences in precision of float numbers
- I have added appropriate lines into resultmap file
- I have modified the script regress.sh to use "case" statement when testing
the hostname. For cygwin the script is called with "i686-pc-cygwin" (on my
machine) as a parameter and this was not catched with the "if" statement.
The check was done for PORTNAME (win) and not HOSTNAME (i.86-pc-cygwin*).
The patch for described modifications is included.
All this modifications can be applied to "current" tree too.
The compilation was done on CygwinB20.1 with gcc 2.95, cygipc library 1.05.
The binaries were able to run also on the newest development snapshot
(2000-03-25).
Dan
Make similar changes to hpux templates. Might want to do the same for
other foo_cc and foo_gcc pairs, but will desist until I hear from
someone who uses those platforms.
and do not arbitrarily pull in CFLAGS instead. This caters to platforms
where the C++ compiler does not like all the same switches the C compiler
wants.
In the event of an elog() while the mode was set to immediate write,
there was no way for it to be set back to the normal delayed write.
The mechanism was a waste of space and cycles anyway, since the only user
was varsup.c, which could perfectly well call FlushBuffer directly.
Now it does just that, and the notion of a write mode is gone.
single integers, and lists of names, without surrounding them with quotes.
Remove all tokens which are defined as operators from ColID and ColLabel
to avoid precedence confusion. Thanks to Tom Lane for catching this.
to next integer. Previously, if selectivity was small, we could compute
very tiny scan cost on the basis of estimating that only 0.001 tuple
would be fetched, which is silly. This naturally led to some rather
silly plans...
Move CREATE FUNCTION/WITH clause to end of statement to get around
shift/reduce conflicts with type names containing "WITH".
Add lots of tokens as allowed ColId's and/or ColLabel's,
so this should be a complete set for the v7.0 release.
We still have an internal limit in the ODBC code of 8 columns per key,
but this should lay the groundwork for resolving that.
Includes reformulated query from Tom Lane.
apparently copied from the makefile for the perl5 interface module,
which needs it for reasons explained in src/interfaces/Makefile.
But none of those reasons apply to plperl.
keys lists of Constraint nodes. This eliminates a type pun that would
probably have caused trouble someday, and eliminates circular references
in the parsetree that were causing trouble now.
Also, change parser's uses of strcasecmp() to strcmp(). Since scan.l
has downcased any unquoted identifier, it is never correct to check an
identifier with strcasecmp() in the parser. For example,
CREATE TABLE FOO (f1 int, UNIQUE("F1"));
was accepted, which is wrong, and xlateSqlFunc did more than it should:
select datetime();
ERROR: Function 'timestamp()' does not exist
(good)
select "DateTime"();
ERROR: Function 'timestamp()' does not exist
(bad)
Clean up grotty coding in them, too. AFAICS from the CVS logs, these
have been broken since Postgres95, so I'm not going to insist on an
initdb to fix them now...
it in a separate object. There's no value in keeping the state separate,
and it creates dangling-pointer problems. Also, remove PQsetenv routines
from public API, until and unless they are redesigned to have a safer
interface. Since they were never part of the documented API before 7.0,
it's unlikely that anyone is calling them.
to avoid undue sensitivity to roundoff error, believe that a zero
or slightly negative range estimate should represent a small
positive selectivity, rather than falling back on a generic default
estimate.
1. C++ style comments in C source for ecpg ( // comment )
2. compiler finds wrong include file extern.h in ecpg/lib/descriptor.c
from
include path instead of workdir (rename it ?)
3. fe-connect getsockopt takes a socklen_t as fifth arg not int (use
SOCKET_SIZE_TYPE instead)
4. char vs unsigned char in psql calls to libpq
5. empty define that results in an empty but terminated line ( ; )
Now for all but point 3 I can supply changes to the
compiler flags, to make the compiler less pedantic.
Or is someone interested in the complications ?
in the meantime can someone apply the attached patch ?
Andreas
use a default value that's fairly small. We were generating a result
of about 0.1, but I think 0.01 is probably better --- want to encourage
use of an indexscan in this situation.
costs using the inner path's parent->rows count as the number of tuples
processed per inner scan iteration. This is wrong when we are using an
inner indexscan with indexquals based on join clauses, because the rows
count in a Relation node reflects the selectivity of the restriction
clauses for that rel only. Upshot was that if join clause was very
selective, we'd drastically overestimate the true cost of the join.
Fix is to calculate correct output-rows estimate for an inner indexscan
when the IndexPath node is created and save it in the path node.
Change of path node doesn't require initdb, since path nodes don't
appear in saved rules.
to simplify constant expressions and expand SubLink nodes into SubPlans
is done in a separate routine subquery_planner() that calls union_planner().
We formerly did most of this work in query_planner(), but that's the
wrong place because it may never see the real targetlist. Splitting
union_planner into two routines also allows us to avoid redundant work
when union_planner is invoked recursively for UNION and inheritance
cases. Upshot is that it is now possible to do something like
select float8(count(*)) / (select count(*) from int4_tbl) from int4_tbl
group by f1;
which has never worked before.
command, the entries in template/.similar can really be regular
expressions. This isn't a new feature, just an observation of what the
code already did.
had already been transformed. This led to failure in examples like
UPDATE table SET fld = (SELECT ...). Repair this, and revise the
comments to explain that transformExpr has to be robust against this
condition. Someday we might want to fix the callers so that
transformExpr is never invoked on its own output, but that someday
is not today.
user, so it doesn't need to be translated from the number to the name.
also ``create database ...'' does not take numbers for the encoding, so
the ENCODING variable does not need to be translated to a number, but left
as the text representation. a patch is supplied to make the changes i
have found to work. i was successful dumping and reloading my database
after these changes.
-
John M. Flinchbaugh
incorrect descriptions of a couple of log-related functions.
I will not force an initdb for this, but log() on a numeric won't
work until you do one...
In function parsing, try for an actual function of the given name and
input types before trying to interpret the function call as a type
coercion request, rather than after. Before, a function that had the
same name as a type and operated on a binary-compatible type wouldn't
get invoked. Also, cross-pollinate between func_select_candidates and
oper_select_candidates to ensure that they use as nearly the same
resolution rules as possible. A few other minor code cleanups too.
problem could be lack of parentheses. This addresses cases like
X UserOp UserOp Y, which will be parsed as (X UserOp) UserOp Y,
whereas what likely was wanted was X UserOp (UserOp Y).
16-Mar-00: trailing + or - is not part of the operator unless the operator
also contains characters not present in SQL92-defined operators. This
solves the 'X=-Y' problem without unduly constraining users' choice of
operator names --- in particular, no existing Postgres operator names
become invalid.
Also, remove processing of // comments, as agreed in the same thread.
running gcc and HP's cc with warnings cranked way up. Signed vs unsigned
comparisons, routines declared static and then defined not-static,
that kind of thing. Tedious, but perhaps useful...
/tmp/trace.out.
However, elog.h uses DEBUG as a log-level flag. As a result, tracing is
turned on even if the libpq++.so is built with DEBUG commented out in
the Makefile.
This patch changes libpq++ to use DEBUGFILE instead (which is not
defined anywhere else).
Oliver Elphick
We probably support a superset of the spec, but I don't have the spec
to confirm this.
Update regression tests to include tests for this format.
Update geometry.out with results from Linux RH 5.2 system
(for last decimal place).
We probably support a superset of the spec, but I don't have the spec
to confirm this.
Update regression tests to include tests for this format.
Fix single-space typo in printed message in regress.sh.
actually a type-coercion problem. If you have a function defined on
class A, and class B inherits from A, then the function ought to work
on class B as well --- but coerce_type didn't know that. Now it does.
mark query as having subselects if a subselect was added from a rule
WHERE condition (as opposed to a rule action). Also, fix adjustment
of varlevelsup so that it actually has some prospect of working when
inserting an expression containing a subselect into a subquery.
small changes in formatting.c code (better memory usage ...etc.) and
better
to_char's cache (will fastly for more to_char()s in one query).
(It is probably end of to_char() development in 7.0 cycle.)
Karel
after trying to resolve the item as an input-column name. This allows us
to be compliant with the SQL92 spec for queries that fall within the spec,
while still accepting the same out-of-spec queries as 6.5 did. You'll only
lose if there is an output column name that is the same as an input
column name, but doesn't refer to the same value. 7.0 will interpret
such a GROUP BY spec differently than 6.5 did. No way around that, because
6.5 was clearly not spec compliant.
CREATE DB/DROP DB. If you didn't think they were wrong, try what
happens when you compile with -DCLOBBER_FREED_MEMORY --- database
name displayed in error messages is trashed, because transaction
abort freed it. Also, remove trailing periods in error messages,
per our prevailing style.
Implement TIME WITH TIME ZONE type (timetz internal type).
Remap length() for character strings to CHAR_LENGTH() for SQL92
and to remove the ambiguity with geometric length() functions.
Keep length() for character strings for backward compatibility.
Shrink stored views by removing internal column name list from visible rte.
Implement min(), max() for time and timetz data types.
Implement conversion of TIME to INTERVAL.
Implement abs(), mod(), fac() for the int8 data type.
Rename some math functions to generic names:
round(), sqrt(), cbrt(), pow(), etc.
Rename NUMERIC power() function to pow().
Fix int2 factorial to calculate result in int4.
Enhance the Oracle compatibility function translate() to work with string
arguments (from Edwin Ramirez).
Modify pg_proc system table to remove OID holes.
(ie, allow rounding to occur at a digit position left of the decimal
point). Apparently this is how Oracle handles it, and there are
precedents in other programming languages as well.
Since we detect oversize tuples elsewhere, I see no reason not to allow
string constants that are 'too long' --- after all, they might never get
stored in a tuple at all.
YY_READ_BUF_SIZE, which turns out to have nothing to do with buffer size.
It's just a totally arbitrary upper limit on how much data myinput() is
asked for at one time.
that the inputs to a given operator can be recursively simplified to
constants, it was evaluating the operator using the op's *original*
(unsimplified) arg list, so that any subexpressions had to be evaluated
again. A constant subexpression at depth N got evaluated N times.
Probably not very important in practical situations, but it made us look
real slow in MySQL's 'crashme' test...
gone, replaced by plain a_expr. The few places where we needed to
distinguish NULL from a_expr are now handled by tests inside the actions
rather than by separate productions. This allows us to accept queries
like 'SELECT 1 + NULL' without requiring parentheses around the NULL.
subplan: do it if subplan has subplans itself, and always do it if the
subplan is an indexscan. (I originally set it to materialize an indexscan
only if the indexqual is fairly selective, but I dunno what I was
thinking ... an unselective indexscan is still expensive ...)
coercion code. I'm beginning to wonder why we have separate candidate
selection routines for functions, operators, and aggregates --- shouldn't
this code all be unified? But meanwhile,
SELECT 'a' LIKE 'a';
finally works; the code for dealing with unknown input types for operators
was pretty busted.
per pghackers discussion around 20-Feb. Also add specific error messages
for unterminated comments and unterminated quoted strings. These things
are nonissues for input coming from psql, but they do matter for input
coming from other front ends.
array. This allows processing of conninfo strings to be made thread-safe,
at the cost of a small memory leak in applications that use
PQconndefaults() and are not updated to free the returned array via
the new PQconninfoFree() function. But PQconndefaults() is probably not
used very much, so this seems like a good compromise.
nodes. The former version failed to check permissions of relations that
were referenced in second and later clauses of UNIONs, and it did not
check permissions of tables referenced via inheritance.
The regression test script runcheck.sh doesn't seem able to
handle the blank line on the end of the resultmap file.
Here's a patch to remove it!!
Keith.
1) adds NetBSD shared lib support on both ELF and a.out platforms
2) replaces "-L$(LIBPQDIR) -lpq" with "$(LIBPQ)" defined in
Makefile.global. This makes it much easier to build stuff in
the source tree after you've already installed the libraries.
3) adds TEMPLATEDIR in Makefile.global that indicates where the
database templates are stored. This separates the template files
from real libraries that are installed in $(LIBDIR).
4) changes include order of <readline/readline.h> and <readline.h>.
The latest GNU readline installs its headers under a readline
subdirectory.
In addition to applying the patch below the following files need to be copied:
backend/port/dynloader:
bsd.h -> netbsd.h
bsd.c -> netbsd.c
include/port:
bsd.h -> netbsd.h
makefiles:
Makefile.bsd -> Makefile.netbsd
It would be great to see this incorporated into the source tree before
the 7.0 release is cut.
Thanks!
-- Johnny C. Lam <lamj@stat.cmu.edu>
Here's a patch to fix the " '.' not allowed in db path" problem I ran into.
I removed '.' from the set of illegial characters, but added backtick. I also
included an explicit test for attempting include a reference to a parent dir.
How that?
Ross
I try change prompt in the psql, but it is set '.' (as '%m') for
non-TCP/IP
connection. This small patch try use uname() information for non-TCP/IP
instead '.'.
Karel
the to_char() source code is large, here are regression tests for
numeric/timestamp/int8 part. It is probably enough test for formatting
code in the formatting.c module. The others (float4/float8/int4) types
share this formatting code and eventual bugs for these types aren't
few probable.
Patch fix timestamp_to_char() for infinity/invalid timestamp too.
Karel
(Subj: [PORTS] initdb problem on NT with 7.0). Since nobody helped me,
I had to find out the reson. The difference between NT and Linux (for
instance) is that "open( path, O_RDWR );" opens a file in text mode. So
sometime less block can be read than required.
I suggest a following patch. BTW the situation appeared before, see
hba.c, pqcomm.c and others.
Alexei Zakharov
when you have networks with the same prefix, but different netmasks.
This is due to the fact that occassionally there is random
(uninitialized?)
data in the extra bits past the point where the netmask cares about
them.
ie (real data from a real live database):
10.0/10 == 00001010.00100000.00100000.00011000
10.0/11 == 00001010.00000000.00000000.00000000
^ Bad data, normally never seen
The v4bitncmp() function was only taking one bit length argument so
it would determine that the networks were different, even though
they really aren't (and the netmask test wouldn't be used). This
ONLY happens if the tuple with the longer bit length is used as the
ip_bits() for the v4bitncmp call AND there happens to be junk data
in place in the shorter tuple. Odd and random, but I saw it happen
a couple times so...
Ryan Mooney
as independent clauses in the grammar. analyze.c takes care of putting
the data where it belongs and complaining about invalid combinations.
Also, make TEMP (and TEMPORARY) non-reserved words.
failures. Fix some outright bugs too, including a reference to
uninitialized memory that would cause failures like this one:
select -('1234567890.1234567'::text);
ERROR: Unable to locate type oid 2139062143 in catalog
such as bpchar(char_expression, N), and pull out the attrtypmod that
the function is coercing to. This allows correct deduction of the
column type in examples such as
CREATE VIEW v AS SELECT f1::char(8) FROM tbl;
Formerly we labeled v's column as char-of-unknown-length not char(8).
Also, this change causes the parser not to insert a redundant length
coercion function if the user has explicitly casted an INSERT or UPDATE
expression to the right length.
fields, nor with bpchar and varchar fields that have typmod -1. The
latter effectively have an unspecified length, so I made them display
as char() and varchar() rather than falsely equating them to char(1)
and varchar(1).
it's a good idea to choose the directory size based on the expected
number of entries. But ShmemInitHash was using a hard-wired constant.
Boo hiss. This accounts for recent report of postmaster failure when
asking for 64K or more buffers.
platform (psql and libpq):
The file "config.h.win32" in the include\ directory (from my patch from
2000-01-18) is missing from the tree. It needs to be put back :-)
The following patch has to be applied in the interfaces\libpq directory.
//Magnus
thinks the connection is idle, the error message is displayed as if
it were a NOTICE. This seems better than dropping the message on
the floor ... particularly if the message is the backend telling us
why it's about to close the connection. The previous behavior was
Backend message type 0x45 arrived while idle
pqReadData() -- backend closed the channel unexpectedly.
which is not real helpful.
as a unary minus operator for numeric. Now that long numeric constants
will get converted to NUMERIC in early parsing, it's essential to have
numeric->int8 conversion to avoid 'can't convert' errors on undecorated
int8 constants. Threw in the rest for completeness while I was in the
area.
I did not force an initdb for this, since the system will still run
without the new pg_proc/pg_operator entries. Possibly I should've.
and produce either FLOAT8 or NUMERIC output depending on whether the
value fits in a float8 or not. This is almost back to the way the
code was before I changed T_Float, but there is a critical difference:
now, when a numeric constant doesn't fit in float8, it will be treated
as type NUMERIC instead of type UNKNOWN.
2. Regression tests fail for types int2 and int4 (which can easily be
fixed by adding entries to resultmap) aswell as float8 and geometry,
where floating point numbers appear to be rounded a little differently
than in your expected results (besides that I also need the positive
zeros file). I'm including a patch for the first 2, but I don't know
whether the latter two are actually a bug in postgres or a bug in the
OS or even allowed difference. I'm including my results for reference.
Rolf Grossmann
tests for the Foreign Key support in 7.0 which
was made against a CVS copy from this
afternoon.
This modifies
src/test/regress/sql/run_check.tests
src/test/regress/sql/alter_table.sql
src/test/regress/expected/alter_table.out
src/test/regress/sql/foreign_key.sql
src/test/regress/expected/foreign_key.out
sszabo@bigpanda.co
integers) to be strings instead of 'double'. We convert from string form
to internal representation only after type resolution has determined the
correct type for the constant. This eliminates loss-of-precision worries
and gets rid of the change in behavior seen at 17 digits with the
previous kluge.
portion of the query result that will be retrieved. As far as I could
tell, the consensus was that we should let the planner do the best it
can with a LIMIT query, and require the user to add ORDER BY if he
wants consistent results from different LIMIT values.
as representing a type coercion request in more cases than we did before.
It will work now whenever no underlying function is required, ie if the
coercion is binary-compatible or if the argument is a previously untyped
string constant. Otherwise, you still need a real function to exist.
represent the result of a binary-compatible type coercion. At runtime
it just evaluates its argument --- but during type resolution, exprType
will pick up the output type of the RelabelType node instead of the type
of the argument. This solves some longstanding problems with dropped
type coercions, an example being 'select now()::abstime::int4' which
used to produce date-formatted output, not an integer, because the
coercion to int4 was dropped on the floor.
agg_select_candidate, which could cause them to keep more candidates
than they should and thus fail to select a single match. I had
previously fixed the identical bug in oper_select_candidate, but
didn't realize that the same error was repeated over here.
Also, repair func_select_candidate's curious notion that it could
scribble on the input type-OID vector. That was causing failure to
apply necessary type coercion later on, leading to malfunction of
examples such as select date('now').
a few bricks shy of a load concerning knowing all the date/time types.
This is real bad because it interferes with func_select_candidate()'s
willingness to disambiguate functions --- func_select_candidate() will
punt unless all the available choices have the same type category.
I think this whole mechanism needs redesigned, but in the meantime
this is a needed patch.
command line processing. As it stood, a bogus PGOPTIONS value from
a client would force a database system restart. Not bad as a denial-
of-service attack...
interpret a column name as an output column alias (targetlist AS name),
ather than a real column name as it ought to. According to the spec,
only ORDER BY should look at output column names. I left in GROUP BY's
willingness to use an output column number ('GROUP BY 2'), even though
this is also contrary to the spec --- again, only ORDER BY is supposed
to accept that. But there is no possible reason to want to GROUP BY
an integer constant, so keeping this old behavior won't break any
SQL-compliant queries. DISTINCT ON will behave the same as GROUP BY.
Change numerology regress test, which depended on the incorrect
behavior.
erroneous expected output for RESET DateStyle: should be ISO now.
Fix run_check.sh so that test postmaster is started with PGDATESTYLE=ISO,
else the horology test won't pass.
variable, instead calling same code in variable.c that is used to parse
SET DATESTYLE. Fix bug: although backend's startup datestyle had been
changed to ISO, 'RESET DATESTYLE' and 'SET DATESTYLE TO DEFAULT' didn't
know about it. For consistency I have made the latter two reset to the
PGDATESTYLE-defined initial value, which may not be the same as the
compiled-in default of ISO.
equivalent now, which should make Windows and Mac clients happier.
Also fix failure to handle SQL comments between segments of a multiline
quoted literal.
appropriate btree three-way comparison routine. Not clear why the
three-way comparison routines were being used in some paths and not
others in btree --- incomplete changes by someone long ago, maybe?
Anyway, this makes for a nice speedup in CREATE INDEX.
selectivity functions and make the r-tree operators use them. The
estimation functions themselves are just stubs, unfortunately, but
perhaps someday someone will make them compute realistic estimates.
Change pg_am so that the optimizer can reliably tell the difference
between ordered and unordered indexes --- before it would think that
an r-tree index can be scanned in '<<' order, which is not right AFAIK.
Repair broken negator links for network_sup and related ops.
Initdb forced. This might be my last initdb force for 7.0 ... hope so
anyway ...
Version: 6.5.3-11
Severity: important
'char' is not a signed type by default on powerpc; therefore a character
can
never be equal to EOF (-1). A patch is attached.
Dan
/--------------------------------\ /--------------------------------\
| Daniel Jacobowitz |__| SCS Class of 2002 |
Implement "date/time grand unification".
Transform datetime and timespan into timestamp and interval.
Deprecate datetime and timespan, though translate to new types in gram.y.
Transform all datetime and timespan catalog entries into new types.
Make "INTERVAL" reserved word allowed as a column identifier in gram.y.
Remove dt.h, dt.c files, and retarget datetime.h, datetime.c as utility
routines for all date/time types.
date.{h,c} now deals with date, time types.
timestamp.{h,c} now deals with timestamp, interval types.
nabstime.{h,c} now deals with abstime, reltime, tinterval types.
Make NUMERIC a known native type for purposes of type coersion. Not tested.
Implement "date/time grand unification".
Transform datetime and timespan into timestamp and interval.
Deprecate datetime and timespan, though translate to new types in gram.y.
Transform all datetime and timespan catalog entries into new types.
Make "INTERVAL" reserved word allowed as a column identifier in gram.y.
Remove dt.h, dt.c files, and retarget datetime.h, datetime.c as utility
routines for all date/time types.
date.{h,c} now deals with date, time types.
timestamp.{h,c} now deals with timestamp, interval types.
nabstime.{h,c} now deals with abstime, reltime, tinterval types.
Make NUMERIC a known native type for purposes of type coersion. Not tested.
Transform datetime and timespan into timestamp and interval.
Deprecate datetime and timespan, though translate to new types in gram.y.
Transform all datetime and timespan catalog entries into new types.
Make "INTERVAL" reserved word allowed as a column identifier in gram.y.
Remove dt.h, dt.c files, and retarget datetime.h, datetime.c as utility
routines for all date/time types.
date.{h,c} now deals with date, time types.
timestamp.{h,c} now deals with timestamp, interval types.
nabstime.{h,c} now deals with abstime, reltime, tinterval types.
Make NUMERIC a known native type for purposes of type coersion. Not tested.
Implement "date/time grand unification".
Transform datetime and timespan into timestamp and interval.
Deprecate datetime and timespan, though translate to new types in gram.y.
Transform all datetime and timespan catalog entries into new types.
Make "INTERVAL" reserved word allowed as a column identifier in gram.y.
Remove dt.h, dt.c files, and retarget datetime.h, datetime.c as utility
routines for all date/time types.
date.{h,c} now deals with date, time types.
timestamp.{h,c} now deals with timestamp, interval types.
nabstime.{h,c} now deals with abstime, reltime, tinterval types.
Make NUMERIC a known native type for purposes of type coersion. Not tested.
accesses versus sequential accesses, a (very crude) estimate of the
effects of caching on random page accesses, and cost to evaluate WHERE-
clause expressions. Export critical parameters for this model as SET
variables. Also, create SET variables for the planner's enable flags
(enable_seqscan, enable_indexscan, etc) so that these can be controlled
more conveniently than via PGOPTIONS.
Planner now estimates both startup cost (cost before retrieving
first tuple) and total cost of each path, so it can optimize queries
with LIMIT on a reasonable basis by interpolating between these costs.
Same facility is a win for EXISTS(...) subqueries and some other cases.
Redesign pathkey representation to achieve a major speedup in planning
(I saw as much as 5X on a 10-way join); also minor changes in planner
to reduce memory consumption by recycling discarded Path nodes and
not constructing unnecessary lists.
Minor cleanups to display more-plausible costs in some cases in
EXPLAIN output.
Initdb forced by change in interface to index cost estimation
functions.
SELECT a FROM t1 tx (a);
Allow join syntax, including queries like
SELECT * FROM t1 NATURAL JOIN t2;
Update RTE structure to hold column aliases in an Attr structure.
Add "SESSION_USER" as SQL92 keyword; equivalent to CURRENT_USER for now.
Implement column aliases (aka correlation names) and more join syntax.
Fix up indenting and tabbing.
where else to mail it. I am the maintainer of unixODBC, and we have a
set of code in our project that started life as the Postgres windows
ODBC driver, which has been ported back to unix. Anyway I have just
fixed a memory leak in the driver, and I cant see any mention of the fix
being done in the main Postgres code, so I thougth I would let you know.
Its in the statement.c module, after the COMMIT statement has been
executed in SC_Execute, the code was
Nick Gorham
this is an old patch which I have already submitted and never seen
in the sources. It corrects the datatype oids used in some iterator
functions. This bug has been reported to me by many other people.
contrib-datetime.patch
some code contributed by Reiner Dassing <dassing@wettzell.ifag.de>
contrib-makefiles.patch
fixes all my contrib makefiles which don't work with some compilers,
as reported to me by another user.
contrib-miscutil.patch
an old patch for one of my old contribs.
contrib-string.patch
a small change to the c-like text output functions. Now the '{'
is escaped only at the beginning of the string to distinguish it
from arrays, and the '}' is no more escaped.
elog-lineno.patch
adds the current lineno of CopyFrom to elog messages. This is very
useful when you load a 1 million tuples table from an external file
and there is a bad value somehere. Currently you get an error message
but you can't know where is the bad data. The patch uses a variable
which was declared static in copy.c. The variable is now exported
and initialized to 0. It is always cleared at the end of the copy
or at the first elog message or when the copy is canceled.
I know this is very ugly but I can't find any better way of knowing
where the copy fails and I have this problem quite often.
plperl-makefile.patch
fixes a typo in a makefile, but the error must be elsewhere because
it is a file generated automatically. Please have a look.
tprintf-timestamp.patch
restores the original 2-digit year format, assuming that the two
century digits don't carry much information and that '000202' is
easier to read than 20000202. Being only a log file it shouldn't
break anything.
Please apply the patches before the next scheduled code freeze.
I also noticed that some of the contribs don't compile correcly. Should we
ask people to fix their code or rename their makefiles so that they are
ignored by the top makefile?
--
Massimo Dal Zotto
a switch statement that has an empty default label. A label of a
switch statement must be followed by a statement (or a label which
is followed by a statement (or a label which ...)).
3. Files include stringinfo.h failed to compile. The macro,
'appendStringInfoCharMacro' is implemented with a '?:' operation
that returns a void expression for the true part and a char expresion
for the false part. Both the true and false parts of the '?:' oper-
ator must return the same type.
Billy G. Allie
am including a patch to get it compile.
changes to psql:
- added less as default pager when compiling on Cygwin
- need to declare "filename_completion_function" because it is not exported
from readline -> added to include/port/win.h
changes to pg_id:
- include of <getopt.h>
- add .exe when installing
I think there is a problem with calling the regress tests on WinNT - it
should be called with PORTNAME not HOST as the parameter to regress.sh or
the check when to add "-h localhost" to psql has to be changed. Now it is
checked against the PORTNAME.
The results of the regress tests were OK with expected failures ;-)
Daniel Horak
and initdb crashs (I set pglib path to PG 6.5.3 directory instead to
7.0 and initdb take this BKI old templates ... (initdb not check
BKI version and BKI files not has any version comments (TODO?))
This patch add to the initdb --show option which show setting of all
initdb's values. It spare developers time if in setting is bug.
Karel
----------------------------------------------------------------------
Karel Zak <zakkr@zf.jcu.cz> http://home.zf.jcu.cz/~zakkr/
The PostgreSQL's to_char() is very compatible with Oracle's to_char
now. I hope that to_char's 3000 rows of source is without bugs, but
will good if anyone test it, for me it works very well :-)
Karel
----------------------------------------------------------------------
Karel Zak <zakkr@zf.jcu.cz> http://home.zf.jcu.cz/~zakkr/
fields in JoinPaths --- turns out that we do need that after all :-(.
Also, rearrange planner so that only one RelOptInfo is created for a
particular set of joined base relations, no matter how many different
subsets of relations it can be created from. This saves memory and
processing time compared to the old method of making a bunch of RelOptInfos
and then removing the duplicates. Clean up the jointree iteration logic;
not sure if it's better, but I sure find it more readable and plausible
now, particularly for the case of 'bushy plans'.
regression tests so I prepared a set of expected
files to make things look OK.
There's also a file to account for minor variations
in the geopmetry output and a resultmap patch to
pull them all together.
With these changes PostgreSQL, from CVS, builds and
regression tests (runcheck) cleanly.
Keith Parks.
nonoverlap_sets() and is_subset() to list.c, where they should have lived
to begin with, and rename to nonoverlap_setsi and is_subseti since they
only work on integer lists.
extracting from an AND subclause just those opclauses that are relevant
for a particular index. For example, we can now consider using an index
on x to process WHERE (x = 1 AND y = 2) OR (x = 2 AND y = 4) OR ...
it seems more suitable for the naming convention in libpq.
New function PQsetClientEncoding added. It makes possible to change
the client encoding on the fly without setting PGCLIENTENCODING.
Added constraint dumping capability to pg_dump (also from Stephan)
Fixed DROP TABLE -> RelationBuildTriggers: 2 record(s) not found for rel
error.
Fixed little error in gram.y I made the last days.
Jan
the cache context, it didn't bother to free the tuple that
CatalogIndexFetchTuple had allocated in the transaction context.
Do enough cache lookups in the same xact, and you start to notice...
syscache and relcache flushes). Relcache entry rebuild now preserves
original tupledesc, rewrite rules, and triggers if possible, so that pointers
to these things remain valid --- if these things change while relcache entry
has positive refcount, we elog(ERROR) to avoid later crash. Arrange for
xact-local rels to be rebuilt when an SI inval message is seen for them,
so that they are updated by CommandCounterIncrement the same as regular rels.
(This is useful because of Hiroshi's recent changes to process our own SI
messages at CommandCounterIncrement time.) This allows simplification of
some routines that previously hacked around the lack of an automatic update.
catcache now keeps its own copy of tupledesc for its relation, rather than
depending on the relcache's copy; this avoids needing to reinitialize catcache
during a cache flush, which saves some cycles and eliminates nasty circularity
problems that occur if a cache flush happens while trying to initialize a
catcache.
Eliminate a number of permanent memory leaks that used to happen during
catcache or relcache flush; not least of which was that catcache never
freed any cached tuples! (Rule parsetree storage is still leaked, however;
will fix that separately.)
Nothing done yet about code that uses tuples retrieved by SearchSysCache
for longer than is safe.
we *always* rebuild, rather than deleting, an invalidated relcache entry
that has positive refcount. Otherwise an SI cache overrun leads to
dangling Relation pointers all over the place!
Initdb help correction
Changed end/abort to commit/rollback and changed related notices
Commented out way old printing functions in libpq
Fixed a typo in alter table / alter column
3 new files and two patches for the plperl subdir.
These changes add the ability for plperl functions
to call 'elog'. It also sets up the frame work to
allow me to add access to the SPI functions.
--
Mark Hollomon
2-Oct-98 or TODO.detail/cnfify) to decide whether we want to reduce
WHERE clause to CNF form, DNF form, or neither. This is a HUGE win.
The heuristic conditions could probably still use a little tweaking to
make sure we don't pick CNF when DNF would be better, or vice versa,
but the risk of exponential explosion in cnfify() is gone. I was able
to run ten-thousand-AND-subclause queries through the planner in a
reasonable amount of time.
SELECT DISTINCT ON (expr [, expr ...]) targetlist ...
and there is a check to make sure that the user didn't specify an ORDER BY
that's incompatible with the DISTINCT operation.
Reimplement nodeUnique and nodeGroup to use the proper datatype-specific
equality function for each column being compared --- they used to do
bitwise comparisons or convert the data to text strings and strcmp().
(To add insult to injury, they'd look up the conversion functions once
for each tuple...) Parse/plan representation of DISTINCT is now a list
of SortClause nodes.
initdb forced by querytree change...
family functions. Contain:
conversion from a datetype to formatted text:
to_char( datetime, text)
to_char( timestamp, text)
to_char( int4, text)
to_char( int8, text)
to_char( float4, text)
to_char( float8, text)
to_char( numeric, text)
vice versa:
to_date ( text, text)
to_datetime ( text, text)
to_timestamp ( text, text)
to_number ( text, text) (convert to numeric)
PostgreSQL to_char is very compatible with Oracle's to_char(), but not
total exactly (now). Small differentions are in number formating. It will
fix in next to_char() version.
! If will this patch aplly to the main tree, must be delete the current
to_char version in contrib (directory "dateformat" and note in contrib's
README), this patch not erase it (sorry Bruce).
The patch patching files:
doc/src/sgml/func.sgml
^^^^^^^^
Hmm, I'm not sure if my English... :( Check it anyone (volunteer)?
Thomas, it is right? SGML is not my primary lang and compile
the current PG docs tree is very happy job (hard variables setting in
docs/sgml/Makefile --> HSTYLE= /home/users/t/thomas/.... :-)
What add any definition to global configure.in and set Makefiles in docs
tree via ./configure?
src/backend/utils/adt/Makefile
src/backend/utils/adt/formatting.c
src/include/catalog/pg_proc.h
src/include/utils/formatting.h
Karel Zak <zakkr@zf.jcu.cz> http://home.zf.jcu.cz/~zakkr/
(ie, WHERE x > lowbound AND x < highbound). It's not very bright yet
but it does something useful. Also, rename intltsel/intgtsel to
scalarltsel/scalargtsel to reflect usage better. Extend convert_to_scalar
to do something a little bit useful with string data types. Still need
to make it do something with date/time datatypes, but I'll wait for
Thomas's datetime unification dust to settle first. Eventually the
routine ought not have any type-specific knowledge at all; it ought to
be calling a type-dependent routine found via a pg_type column; but
that's a task for another day.
an attribute of a tuple previously fetched with SearchSysCacheTuple.
This avoids a lot of redundant cache lookups, particularly in selfuncs.c.
Also, remove SearchSysCacheStruct, which was unused and grotty.
pghackers discussion of 5-Jan-2000. The amopselect and amopnpages
estimators are gone, and in their place is a per-AM amcostestimate
procedure (linked to from pg_am, not pg_amop).
have the rl_completion_append_character variable. The tab completion
behavior doesn't seem to be quite perfect in that situation, but it's
better than failing to build at all...
does not end with a newline. I don't think this explains the recent
complaints, since this bug existed in 6.5 (and probably long before).
But might as well fix it now that I see it.
fact the same, so I suggest they could be the same file say
geometry-positive-zeros.out, as the main difference seems to be not printing
eg. (0,-0). In src/test/regress/expected, I propose
rm int2-i386-netbsd.out int4-i386-netbsd.out
mv geometry-hppa1.1.out geometry-positive-zeros.out
rm geometry-hppa2.0.out geometry-i386-netbsd.out
and the following patch to resultmap. I have only tested the netbsd results
on i386, but think that in all probability the differences will be the same
for other ports. If it turns out not to be the case, at least we might find
out.
Patrick Welche
Attached is a small fix for a stupid mistake I made in comment.c
- an attempt to drop a non-existent comment would dump core :-(.
Sometimes, I'm as sharp as a marble.
Sorry,
Mike Mascari
compiler warnings caused by lack of extern declarations in extern.h.
I believe the remaining gcc warnings here would go away if the ecpg
grammar could be tweaked so it doesn't use REJECT ...
that kept me from making perl secure.
Attached is uuencoded tarball to add PL/perl
to postgresql.
Things I know don't work.
-- triggers
-- SPI
The README file has a _VERY_ short tutorial.
Mark Hollomon
allows casts without specific length requirements to continue to work
as they did before; that is, x::char will not truncate the value of x,
whereas x::char(1) will. Likewise for NUMERIC precision/scale.
The column length defaults of char(1) and numeric(30,6) are now inserted
in analyze.c's processing of CREATE TABLE.
from a constraint condition does not violate the constraint (cf. discussion
on pghackers 12/9/99). Implemented by adding a parameter to ExecQual,
specifying whether to return TRUE or FALSE when the qual result is
really NULL in three-valued boolean logic. Currently, ExecRelCheck is
the only caller that asks for TRUE, but if we find any other places that
have the wrong response to NULL, it'll be easy to fix them.
Here is a patch to bring both libpq and psql to a state where it compiles on
win32 (native) again. A lot of things have changed, and I have not been able
to keep up with them all, so it has been broken for quite a while.
After this patch, at least it compiles. It also talks "basic talk" to the
server, but I have not yet tested all things. Sending queries, and using
e.g. \d or \dt works fine. The rest will have to be tested further.
It also bumps the version on libpq.dll to 7.0.
Everything should be enclosed in #ifdef WIN32, unless I have missed
something. Except for one or maybe two places where I have moved a #include
that should not be used on win32 from the "global area" into a "#ifndef
WIN32 area".
//Magnus
Attached is a patch which patches cleanly against the Sunday afternoon
snapshot. It modifies pg_dump to dump COMMENT ON statements for
user-definable descriptions. In addition, it also modifies comment.c so
that the operator behavior is as Peter E. would like: a comment on an
operator is applied to the underlying function.
Thanks,
Mike Mascari
is considerably more robust and accurate than it used to be.
Also, get rid of numeric's private allocation freelist, which is no longer
a win since Jan rewrote palloc.
SQL cast constructs can be performed during expression transformation
instead of during parsing. This allows constructs like x::numeric(9,2)
and x::int2::float8 to behave as one would expect.
read is reused for successive attributes, instead of being deleted and
recreated from scratch for each value read in. This reduces palloc/pfree
overhead a lot. COPY IN still seems to be noticeably slower than it was
in 6.5 --- we need to figure out why. This change takes care of the only
major performance loss I can see in copy.c itself, so the performance
problem is at a lower level somewhere.
by creating a race condition. It wasn't waiting for select() to say
write-ready immediately after connect, which meant that you might get
an unhelpful 'broken pipe' error message if connect failed, rather than
the intended error message.
CommandCounterIncrement to make new relation visible before trying to
parse/deparse the expressions. Also, eliminate unnecessary
setheapoverride calls in AddNewAttributeTuples.
oidvector/int2vector. pg_dump code was assuming that it would see
exactly FUNC_MAX_ARGS integers in the string returned by the backend.
That's no longer true. (Perhaps that change wasn't such a good idea
after all --- will it break any other applications??)
functions, which would lead to trouble with datatypes that paid attention
to the typelem or typmod parameters to these functions. In particular,
incorrect code in pg_aggregate.c explains the platform-specific failures
that have been reported in NUMERIC avg().
- Prevent permissions on indexes
- Instituted --enable-multibyte option and tweaked the MB build process where necessary
- initdb prompts for superuser password
* Let unprivileged users change their own passwords.
* The password is now an Sconst in the parser, which better reflects its text datatype and also
forces users to quote them.
* If your password is NULL you won't be written to the password file, meaning you can't connect
until you have a password set up (if you use password authentication).
* When you drop a user that owns a database you get an error. The database is not gone.
in libpq --- mostly, poor response to error conditions. You now actually
get to see the postmaster's 'The Data Base System is starting up' message,
which you didn't before. I suspect the SSL code is still broken though.
choke on relation or attribute names containing spaces, quotes, or other
special characters. This fixes a TODO item. It also forces initdb,
since stored rule strings change.
errors. VACUUM normally compacts the table back-to-front, and stops
as soon as it gets to a page that it has moved some tuples onto.
(This logic doesn't make for a complete packing of the table, but it
should be pretty close.) But the way it was checking whether it had
got to a page with some moved-in tuples was to look at whether the
current page was the same as the last page of the list of pages that
have enough free space to be move-in targets. And there was other
code that would remove pages from that list once they got full.
There was a kluge that prevented the last list entry from being
removed, but it didn't get the job done. Fixed by keeping a separate
variable that contains the largest block number into which a tuple
has been moved. There's no longer any need to protect the last element
of the fraged_pages list.
Also, fix NOTICE messages to describe elapsed user/system CPU time
correctly.
Instead of hard-wiring one result file per platform, there is a map file
'resultmap' that says which one to use --- a lot like template/.similar.
I have only created entries in resultmap for my own platform (HPUX) so
far; feel free to add lines for other platforms.
which is broken in some weird way that I don't understand. I think it
may be exposing a bug in the new psql --- for one thing, I get different
results when I run psql by hand than the regress script gets. What
the heck???
quote_postgres(...) in ecpglib.c.
The code in CVS reads:
quote_postgres(char *arg, int lineno)
{
char *res = (char *) ecpg_alloc(2 * strlen(arg) + 3, lineno);
int i,
ri = 0;
if (!res)
return (res);
res[ri++] = '\'';
for (i = 0, ri=0; arg[i]; i++, ri++)
{
switch (arg[i])
{
case '\'':
res[ri++] = '\'';
break;
case '\\':
res[ri++] = '\\';
break;
default:
;
}
The problem here is that ri is reset to 0, thus overwriting the initial
quote.
Stephen Birch
if presented an uninitialized (all zeroes) page. The system no longer
crashes hard if an all-zeroes page is present in a relation. There seem
to be some boundary conditions where a page will be appended to a relation
and zeroed, but its page header is never initialized; until we can track
down and fix all of those, robustness seems like a good idea.
Also, clean up some obsolete and downright wrong comments.
1) datetime_pl_span() added the seconds field before adding the months
field. This lead to erroneous results for e.g.
select datetime '1999-11-30' + timespan '1 mon - 1 sec';
Reverse the order of operations to add months first.
2) tm2timespan() did all intermediate math as integer, converting to double
at the very end. This resulted in hidden overflows when given very large
integer days, hours, etc. For example,
select '74565 days'::timespan;
produced the wrong result. Change code to ensure that doubles are used
for intermediate calculations.
Thanks to Olivier PRENANT <ohp@pyrenet.fr> and
Tulassay Zsolt <zsolt@tek.bke.hu> for problem reports and to Tom Lane for
accurate analyses.
during InitProcessingMode and the CurrentTransactionState was neither
TRANS_DEFAULT nor TRANS_DISABLED. Unfortunately, after someone's recent
change to start the transaction manager earlier in startup than it used
to be started, that caused an abort() and consequent database system
reset on quite harmless errors (such as rejecting an invalid user name!).
As far as I can see, the test on CurrentTransactionState was completely
useless anyway, so I've removed it.
relcache entry no longer leaks a small amount of memory. index_endscan
now releases all the memory acquired by index_beginscan, so callers of it
should NOT pfree the scan descriptor anymore.
I finally got around to schlepping through pg_dump, to finish what I started
about three months (or more) ago. Attached is a gzipped diff file to apply
in the bin/pg_dump directory. This should remove all string length
dependencies, except one, which I'm working on. It has been through some
rudimentary unit testing, but that's about it, so if any of you would give
it a more strenuous run-through, I'd be grateful for the feedback.
Cheers...
Ansley, Michael
SELECT null::text;
SELECT int4fac(null);
work as expected now. In some cases a NULL must be surrounded by
parentheses:
SELECT 2 + null; fails
SELECT 2 + (null); OK
This is a grammatical ambiguity that seems difficult to avoid. Other
than that, NULLs seem to behave about like you'd expect. The internal
implementation is that NULL constants are typed as UNKNOWN (like
untyped string constants) until the parser can deduce the right type.
Locate path of postmaster in a portable way (stolen from initdb)
Add postmaster.opts.default.sample which should be copied into
$PGLIB in the installtion process. Also, it will be installed into
$PGDATA while initdb is running.
a ".pgc " extension. The second patch fixes a coredump when there is
more than one input file (in that case, cur and types were not set to
NULL before processing the second f ile)
The patch below modifies the accepted grammar of ecpg to accept
FETCH [direction] [amount] cursor name
i.e. the IN|FROM clause becomes optional (as in Oracle and Informix).
This removes the incompatibility mentioned in section "Porting From
Other RDBMS Packages" p169, PostgreSQL Programmer's Guide. The grammar
is modified in such a way as to avoid shift/reduce conflicts. It does
not accept the statement "EXEC SQL FETCH;" anymore, as the old grammar
did (this seems to be a bug of the old grammar anyway).
This patch cleans up the handling of space characters in the scanner;
some patte rns require \n to be in {space}, some do not. A second fix is
the handling of cpp continuati on lines; the old pattern did not match
these. The parser is patched to fix an off-by-one error in the #line
directives. The pa rser is also enhanced to report the correct location
of errors in declarations in the "E XEC SQL DECLARE SECTION". Finally,
some right recursions in the parser were replaced by left-recursions.
This patch adds preprocessor directives to ecpg; in particular
EXEC SQL IFDEF, EXEC SQL IFNDEF, EXEC SQL ELSE, EXEC SQL ELIF and EXEC SQL ENDIF
"EXEC SQL IFDEF" is used with defines made with "EXEC SQL DEFINE" and
defines, specified on the command line with -D. Defines, specified on
the command line are persistent across multiple input files. Defines can
be nested up to a maximum level of 128 (see patch). There is a fair
amount of error checking to make sure directives are matched properly. I
need preprocessor directives for porting code, that is written for an
Informix database, to a PostgreSQL database, while maintaining
compatibility with the original code. I decided not to extend the
already large ecpg grammar. Everything is done in the scanner by adding
some states, e.g. to skip all input except newlines and directives. The
preprocessor commands are compatible with Informix. Oracle uses a cpp
replacement.
Rene Hogendoorn
with DEC C.
DEC C doesn't handle double values greater than DBL_MAX, but some
PostgreSQL geo functions assign greater than DBL_MAX values to some vars
in some special cases - that couses SIGFPE. I dunno if that is the only place
to fix to work well with DEC C.
Kirill Nosov.
rather than returning a NaN for bogus input to pow(). Namely, HPUX 10.20.
I think this is sufficient evidence for what I thought all along, which
is that the float.c code *must* look at errno whether finite() exists or
not.
> > for them to actually set out and do it. Many new users are
> > of the not-so-knowledgable variety, and shell scripting isn't
> > something they want to undertake.
>
> Can someone modify the vacuumdb shell script to do that?
i tried it... it seems to work
neko@kredit.sth.sz
initdb. No more obscure dependencies on environment variables or paths.
It
now finds the templates and the right postgres itself (with cmd line
options as fallback). It also no longer depends on $USER (su safe), and
doesn't advertise that --username allows you to install the db as a
different user, since that doesn't work anyway. Also, recovery and
cleanup
on all errors. Consistent options, clearer documentation.
Please take a look at this and adopt it if you feel it's safe enough. I
have simulated all the stupid circumstances I could think of, but you
never know with shell scripts.
Oh yeah, you can give the postgres user a default password now.
--
Peter Eisentraut Sernanders väg 10:115
I sending promised patch with:
* getopt_long() - for pg_dump (portable)
* and "Usage: " changes in scripts in src/bin/
- this changes are cosmetic only, not change any
feature ...etc.
All PostgreSQL routines (scripts) support now long options and
help's output is alike for all scripts and all support -? or --help.
Karel
Karel Zak <zakkr@zf.jcu.cz> http://home.zf.jcu.cz/~zakkr/
>go about this. That will risk breaking existing applications that use
>those names as column names.
>
>It should actually almost work to write sq.nextval as things stand,
>because Postgres has for a long time considered table.function and
>function(table) to be interchangeable notations for certain kinds of
>functions. nextval doesn't seem to be one of that kind of function,
>at the moment. I'd suggest leaving the grammar as it was, and taking a
>look at ParseFuncOrColumn in parse_func.c to see if you can't persuade
>it to accept the sequence functions in that style.
OK, good point. I tried to implement it somewhere else and ended up
extending transformAttr. Attached you'll find the patch.
Jeroen van Vianen
didn't have time for documentation yet, but I'll write some. There are
still some things to work out what happens when you alter or drop users,
but the group stuff in and by itself is done.
--
Peter Eisentraut Sernanders väg 10:115
equality. The lobits macro is wrong and extracts the wrong set of
bits out of the structure.
To exhibit the problem:
select '000000:000000'::macaddr = '000000:110000'::macaddr ;
?column?
--------
t
(1 row)
Daniel Boyd
triggered
> function now returns the right datatype.
Oops, I got crossed up with Jan's improvements. Ignore this.
--
Peter Eisentraut Sernanders väg 10:115
peter_e@gmx.net 75262 Uppsala
anywhere from zero to two TODO items.
* Allow flag to control COPY input/output of NULLs
I got this:
COPY table .... [ WITH NULL AS 'string' ]
which does what you'd expect. The default is \N, otherwise you can use
empty strings, etc. On Copy In this acts like a filter: every data item
that looks like 'string' becomes a NULL. Pretty straightforward.
This also seems to be related to
* Make postgres user have a password by default
If I recall this discussion correctly, the problem was actually that the
default password for the postgres (or any) user is in fact "\N", because
of the way copy is used. With this change, the file pg_pwd is copied out
with nulls as empty strings, so if someone doesn't have a password, the
password is just '', which one would expect from a new account. I don't
think anyone really wants a hard-coded default password.
Peter Eisentraut Sernanders väg 10:115
* Document/trigger/rule so changes to pg_shadow recreate pg_pwd
I did it with a trigger and it seems to work like a charm. The function
that already updates the file for create and alter user has been made a
built-in "SQL" function and a trigger is created at initdb time.
Comments around the pg_pwd updating function seem to be worried about
this
routine being called concurrently, but I really don't see a reason to
worry about this. Verify for yourself. I guess we never had a system
trigger before, so treat this with care, and feel free to adjust the
nomenclature as well.
--
Peter Eisentraut Sernanders väg 10:115
at all, and because of shell quoting rules this can't be fixed, so I put
in error messages to that end.
Also, calling create or drop database in a transaction block is not so
good either, because the file system mysteriously refuses to roll back rm
calls on transaction aborts. :) So I put in checks to see if a transaction
is in progress and signal an error.
Also I put the whole call in a transaction of its own to be able to roll
back changes to pg_database in case the file system operations fail.
The alternative location issues I posted recently were untouched, awaiting
the outcome of that discussion. Other than that, this should be much more
fool-proof now.
The docs I cleaned up as well.
Peter Eisentraut Sernanders väg 10:115
yet, but at least we can give a better error message:
regression=> select count(distinct f1) from int4_tbl;
ERROR: aggregate(DISTINCT ...) is not implemented yet
instead of 'parser: parse error at or near distinct'.
I was able to crash postgres 6.5.3 when I did an 'alter user' command.
After I started a debugger I found the problem in the timezone handling
of
datetime (my Linux box lost its timezone information, that's how the
problem occurred).
Only 7 bytes are reserved for the timezone, without checking for
boundaries.
Attached is a patch that fixes this problem and emits a NOTICE if a
timezone is encountered that is longer than MAXTZLEN bytes, like this:
Jeroen van Vianen
against the sources from one hour ago and contain all the portable and
up
to date stuff.
A few other CVS "householding" things you might want to take care of:
* Remove the src/bin/cleardbdir directory
* Remove the file src/bin/psql/sql_help.h from the repository, as it is
a derived file and is build by the release_prep.
Peter Eisentraut
(which are palloc'd) instead of DLLists (which are malloc'd). Not very
significant, since this routine seldom has anything useful to do, but
a leak is a leak...
them into the scripts dir. I also added a --list option to show already
installed languages.
This whole moving and renaming totally confused CVS and my checked out
copy got completely fried last night. When you apply the source patch,
please make sure that all the directories src/bin/{create|destroy}* as
well as vacuumdb, cleardbdir are gone and that all the scripts (7) are
in
scripts/.
Meanwhile I am still puzzled about what happened with the docs patch.
Because I don't know what you got now, the second attachment contains
the
files
ref/allfiles.sgml
ref/commands.sgml
ref/createlang.sgml
ref/droplang.sgml
doc/src/sgml/Makefile
Peter Eisentraut Sernanders väg 10:115
parameter in some flavor of Unix, but Linux, HPUX, and SunOS all say
it's int. For now I'm just going to make it int so that I can compile.
If the other way is actually necessary on some Unix somewhere, I guess
we will need a configure test...
This one should work much better than the one I sent in previously. The
functionality is the same, but the patch was missing one file resulting
in
the compilation failing. The docs also received a minor fix.
Peter Eisentraut Sernanders väg 10:115
PQconnectStart
PQconnectPoll
PQresetStart
PQresetPoll
PQsetenvStart
PQsetenvPoll
PQsetenvAbort
and brings into the published interface
PQsetenv.
The first four are asynchronous analogues of PQconnectdb and PQreset -
they allow an application to connect to the DB without blocking on
remote I/O.
The PQsetenv functions perform an environment negotiation with the
server.
Internal to libpq, pqReadReady and pqWriteReady have been made available
across the library (they were previously static functions inside
fe-misc.c). A lot of internal rearrangement has been necessary to
support these changes.
The API documentation has been updated also.
Caveats:
o The Windows code does not default to using non-blocking sockets,
since I have no documentation: Define WIN32_NON_BLOCKING_CONNECTIONS to
do that.
o The SSL code still blocks.
Ewan Mellor.
You have CommLog and Debug enabled
You encounter in error in any operation (SQLConnect/SQLExec).
Previously, the extra logging didn't check for NULL pointers
when trying to print some of the strings- the socket error
message could frequently be NULL by design (if there was no socket
error)
and Solaris does not handle NULLS passed to things like printf
("%s\n",string);
gracefully.
This basically duplicates the functionality found in Linux where passing
a null pointer
to printf prints "(NULL)". No very elegant, but the logging is for debug
only anyway.
Dirk Niggemann
table owner in order to vacuum a table. This is mainly to prevent
denial-of-service attacks via repeated vacuums. Allow VACUUM to gather
statistics about system relations, except for pg_statistic itself ---
not clear that it's worth the trouble to make that case work cleanly.
Cope with possible tuple size overflow in pg_statistic tuples; I'm
surprised we never realized that could happen. Hold a couple of locks
a little longer to try to prevent deadlocks between concurrent VACUUMs.
There still seem to be some problems in that last area though :-(
expressions were written without spaces between operators and operands.
Problem was that something like "if new.f1=new.f2 then" would be translated
to "if $1=$2 then", and the Postgres lexer would tokenize that the wrong
way. Fix is to emit spaces around $paramno constructs to ensure they are
seen as separate tokens.
parallel --- and, not incidentally, removing a common reason for needing
manual cleanup by the DB admin after a crash. Remove initial global
delete of pg_statistics rows in VACUUM ANALYZE; this was not only bad
for performance of other backends that had to run without stats for a
while, but it was fundamentally broken because it was done outside any
transaction. Surprising we didn't see more consequences of that.
Detect attempt to run VACUUM inside a transaction block. Check for
query cancel request before starting vacuum of each table. Clean up
vacuum's private portal storage if vacuum is aborted.
By dropping stats rows here, we eliminate the need for VACUUM to do a
wholesale remove of stats rows. Before, pg_statistics was wiped clean
at the start of VACUUM, ensuring poor planning results for any backends
running in parallel until VACUUM got around to rebuilding the stats for
the relations they are accessing.
rate
it's better than what used to be there.
* Does proper SQL "host variable" substitution as pointed out by Andreas
Zeugwetter (thanks): select * from :foo; Also some changes in how ':'
and ';' are treated (escape with \ to send to backend). This does
_not_
affect the '::' cast operator, but perhaps others that contain : or ;
(but there are none right now).
* To show description with a <something> listing, append '?' to command
name, e.g., \df?. This seemed to be the convenient and logical
solution.
Or append a '+' to see more useless information, e.g., \df+.
* Fixed fflush()'ing bug pointed out by Jan during the regression test
discussion.
* Added LastOid variable. This ought to take care of TODO item "Add a
function to return the last inserted oid, for use in psql scripts"
(under CLIENTS)
E.g.,
insert into foo values(...);
insert into bar values(..., :LastOid);
\echo $LastOid
* \d command shows constraints, rules, and triggers defined on the table
(in addition to indices)
* Various fixes, optimizations, corrections
* Documentation update as well
Note: This now requires snprintf(), which, if necessary, is taken from
src/backend/port. This is certainly a little weird, but it should
suffice
until a source tree cleanup is done.
Enjoy.
--
Peter Eisentraut Sernanders väg 10:115
Make all system indexes unique.
Make all cache loads use system indexes.
Rename *rel to *relid in inheritance tables.
Rename cache names to be clearer.
for the case of errors in backend startup, and proc_exit's method for
coping with errors during proc_exit was *completely* busted. Fixed per
discussions on pghackers around 11/6/99.
it wants to release. This leads to a race condition: does the backend
that's trying to flush the buffer do so before the one that's deleting the
relation does so? Usually no problem, I expect, but on occasion this could
lead to hard-to-reproduce complaints from md.c, especially mdblindwrt.
returns a list of RelOptInfos, eliminating the need for static state
in index_info. That static state was a direct cause of coredumps; if
anything decided to elog(ERROR) partway through an index_info search of
pg_index, the next query would try to close a scan pointer that was
pointing at no-longer-valid memory. Another example of the reasons to
avoid static state variables...
of the index it wants to destroy. This ensures that no other backend is
actively scanning or updating that index. Getting exclusive access on
the index alone is NOT sufficient, because the executor is rather
cavalier about getting locks on indexes --- see ExecOpenIndices().
It might be better to grab index locks in the executor, but I'm not
sure the extra lockmanager traffic is really worth it just to make
index_destroy cleaner.
(whoever thought world-writable files were a good default????). Modify
the pg_pwd code so that pg_pwd is created with 600 permissions. Modify
initdb so that permissions on a pre-existing PGDATA directory are not
blindly accepted: if the dir is already there, it does chmod go-rwx
to be sure that the permissions are OK and the dir actually is owned
by postgres.
inval.c thought it could safely use the catcache to look up the OIDs of
system relations. Not good, considering that inval.c could be called
during catcache loading, if a shared-inval message arrives. Rip out the
lookup logic and instead use the known OIDs from pg_class.h.
table defaults or rules: translate them to a function call so that
parse_coerce doesn't reduce them to a date or time constant immediately.
Also, eliminate a lot of redundancy in the expression grammar by
defining a new nonterminal com_expr, which contains all the productions
that can be shared by a_expr and b_expr.
Warn_restart has been set by the backend main loop. This means that
elog(ERROR) or elog(FATAL) in the postmaster or during backend startup
now have well-defined behavior: proc_exit() rather than coredump.
In the case of elog() inside the postmaster, I think that proc_exit()
is probably not enough --- don't we want our child backends to be
forced to quit too? But I don't understand Vadim's recent changes in
this area, so I'll leave it to him to look over and tweak if needed.
subselects can only appear on the righthand side of a binary operator.
That's still true for quantified predicates like x = ANY (SELECT ...),
but a subselect that delivers a single result can now appear anywhere
in an expression. This is implemented by changing EXPR_SUBLINK sublinks
to represent just the (SELECT ...) expression, without any 'left hand
side' or combining operator --- so they're now more like EXISTS_SUBLINK.
To handle the case of '(x, y, z) = (SELECT ...)', I added a new sublink
type MULTIEXPR_SUBLINK, which acts just like EXPR_SUBLINK used to.
But the grammar will only generate one for a multiple-left-hand-side
row expression.
circumstances:
=> select * from foo\x\t\pset border 0 \p\g\\select * from bar;
Also the release prep update so the sql_help.h is generated before
packaging.
Peter.
nulls with non-nulls using proper three-valued boolean logic. Also clean
up ExecQual to make it clearer that ExecQual *does* follow the SQL spec
for boolean nulls. See '[BUGS] (null) != (null)' thread around 10/26/99
for more detail.
* Add use of 'const' for varibles in source tree
(which is misspelled, btw.)
I went through the front-end libpq code and did so. This affects in
particular the various accessor functions (such as PQdb() and
PQgetvalue()) as well as, by necessity, the internal helpers they use.
I have been really thorough in that regard, perhaps some people will find
it annoying that things like
char * foo = PQgetvalue(res, 0, 0)
will generate a warning. On the other hand it _should_ generate one. This
is no real compatibility break, although a few clients will have to be
fixed to suppress warnings. (Which again would be in the spirit of the
above TODO.)
In addition I replaced some int's by size_t's and removed some warnings
(and generated some new ones -- grmpf!). Also I rewrote PQoidStatus (so it
actually honors the const!) and supplied a new function PQoidValue that
returns a proper Oid type. This is only front-end stuff, none of the
communicaton stuff was touched.
The psql patch also adds some new consts to honor the new libpq situation,
as well as fixes a fatal condition that resulted when using the -V
(--version) option and there is no database listening.
So, to summarize, the psql you should definitely put in (with or without
the libpq). If you think I went too far with the const-mania in libpq, let
me know and I'll make adjustments. If you approve it, I will also update
the docs.
-Peter
--
Peter Eisentraut Sernanders vaeg 10:115
Now indexes of pg_class and pg_type are unique indexes
and guarantee the uniqueness of correponding attributes.
heap_create() was changed to take another boolean parameter
which allows to postpone the creation of disk file.
The name of rd_nonameunlinked was changed to rd_unlinked.
It is used generally(not only for noname relations) now.
Requires initdb.
Apparently, back in the dim reaches of prehistory, the parser couldn't
be trusted to label Const nodes with the correct constbyval value ...
and someone preferred to patch around this in copyObject rather than
fix the problem at the source. The problem is long gone, but the hack
lingered on. Until now.
in the TupleDesc that the caller already has (for call from ExecMain) or
can make just as easily as ExecInitJunkFilter() can (for call from
ExecAppend). Also, don't bother to build a junk filter for an INSERT
operation that doesn't actually need one, which is the normal case.
In particular, don't bother to look up type information for attributes
where we're not actually going to use it, and avoid copying entire tlist
structure when it's not necessary.
during initial run formation by keeping both current run and next-run
tuples in the same heap (yup, Knuth is smarter than I am). And, during
merge passes, make use of available sort memory to load multiple tuples
from any one input 'tape' at a time, thereby improving locality of
access to the temp file.
before calling execProject, when the outerPlan has returned zero tuples.
I took this out under the mistaken impression that the input tuple
couldn't be referenced by execProject if we weren't in GROUP BY mode.
But it can, if we're in an UPDATE or DELETE...
One, it now returns the previous hook. That way people don't have to dig
around in libpq-int.h for that information anymore. It previously
returned void, so there should be no incompatibilities.
Second, you cannot set the callback to NULL anymore. (Of course you can
still call it with NULL just to get the current hook.) The way libpq uses
the callback pointer, having a NULL there wasn't very healthy.
Peter Eisentraut
The following patch extends the COMMENT ON functionality to the
rest of the database objects beyond just tables, columns, and views. The
grammer of the COMMENT ON statement now looks like:
COMMENT ON [
[ DATABASE | INDEX | RULE | SEQUENCE | TABLE | TYPE | VIEW ] <objname>
|
COLUMN <relation>.<attribute> |
AGGREGATE <aggname> <aggtype> |
FUNCTION <funcname> (arg1, arg2, ...) |
OPERATOR <op> (leftoperand_typ rightoperand_typ) |
TRIGGER <triggername> ON relname>
Mike Mascari
(mascarim@yahoo.com)
eliminating some wildly inconsistent coding in various parts of the
system. I set MAXPGPATH = 1024 in config.h.in. If anyone is really
convinced that there ought to be a configure-time test to set the
value, go right ahead ... but I think it's a waste of time.
when an initdb-forcing change has been applied within a development cycle.
PG_VERSION serves this purpose for official releases, but we can't bump
the PG_VERSION number every time we make a change to the catalogs during
development. Instead, increase the catalog version number to warn other
developers that you've made an incompatible change. See my mail to
pghackers for more info.
fix recently applied to backend's lexer). I see that YY_USES_REJECT
still gets defined for this lexer, which means it's going to have trouble
parsing really long tokens. Not sure if it's worth doing anything about
that or not; I don't have the interest right now to understand why
ecpg's additions to the syntax cause this problem...
This patch fix a TODO list item.
* require SELECT DISTINCT target list to have all ORDER BY columns
example
ogawa=> select distinct x from t1 order by y;
ERROR: ORDER BY columns must appear in SELECT DISTINCT target list
---
Atsushi Ogawa
boundary-condition bug in myinput() which caused flex scanner to fail
on tokens larger than a bufferload. Turns out flex doesn't want null-
terminated input ... and if it gives you a 1-character buffer, you'd
better supply a character, not a null, lest you be thought to be
reporting end of input.
proc_exit time. I discovered that if the frontend closes the connection
when you're inside a transaction block, there is nothing ensuring that
temp files go away ... I wonder whether proc_exit ought to try to do an
explicit transaction abort?
a generalized module 'tuplesort.c' that can sort either HeapTuples or
IndexTuples, and is not tied to execution of a Sort node. Clean up
memory leakages in sorting, and replace nbtsort.c's private implementation
of mergesorting with calls to tuplesort.c.
recycle storage within sort temp file on a block-by-block basis. This
reduces peak disk usage to essentially just the volume of data being
sorted, whereas it had been about 4x the data volume before.
>From the ORACLE 7 SQL Language Reference Manual:
-----------------------------------------------------
COMMENT
Purpose:
To add a comment about a table, view, snapshot, or
column into the data dictionary.
Prerequisites:
The table, view, or snapshot must be in your own
schema
or you must have COMMENT ANY TABLE system privilege.
Syntax:
COMMENT ON [ TABLE table ] |
[ COLUMN table.column] IS 'text'
You can effectively drop a comment from the database
by setting it to the empty string ''.
-----------------------------------------------------
Example:
COMMENT ON TABLE workorders IS
'Maintains base records for workorder information';
COMMENT ON COLUMN workorders.hours IS
'Number of hours the engineer worked on the task';
to drop a comment:
COMMENT ON COLUMN workorders.hours IS '';
The current patch will simply perform the insert into
pg_description, as per the TODO. And, of course, when
the table is dropped, any comments relating to it
or any of its attributes are also dropped. I haven't
looked at the ODBC source yet, but I do know from
an ODBC client standpoint that the standard does
support the notion of table and column comments.
Hopefully the ODBC driver is already fetching these
values from pg_description, but if not, it should be
trivial.
Hope this makes the grade,
Mike Mascari
(mascarim@yahoo.com)
BufFile so that it handles multi-segment temporary files transparently.
This allows sorts and hashes to work with data exceeding 2Gig (or whatever
the local limit on file size is). Change psort.c to use relative seeks
instead of absolute seeks for backwards scanning, so that it won't fail
when the data volume exceeds 2Gig.
I have changed a bit the makefiles for the win32 port - the *.def files
(created when building shared libraries) are now clean from
Makefile.shlib.
I have also removed "-g" from CFLAGS in the "cygwin32" template - it can
be
enabled when running configure.
Dan
database, but they get truncated at the first NUL by lo_read
when they are read back. The reason for this is that lo_read in
Pg.xs is using the default:
OUTPUT:
RETVAL
buf
which uses C's strlen() to work out the length of the scalar.
The code ought to read something more like:
OUTPUT:
RETVAL
buf sv_setpvn((SV*)ST(2), buf, RETVAL);
I am not sure if this needs to be done on both lo_read methods
in this file, but I changed both and have not since had any
problems with truncated BLOBs.
Douglas Thomson <dougt@mugc.cc.monash.edu.au>
Cygwin snapshots (tested on 990115 which is recommended to use - it fixes
some errors in B20.1)
And I have another patch for including <sys/ipc.h> before <sys/sem.h> in
backend/storage/lmgr/proc.c - it is required due the design of cygipc
headers
Dan
mentioned in FROM but not elsewhere in the query: such tables should be
joined over anyway. Aside from being more standards-compliant, this allows
removal of some very ugly hacks for COUNT(*) processing. Also, allow
HAVING clause without aggregate functions, since SQL does. Clean up
CREATE RULE statement-list syntax the same way Bruce just fixed the
main stmtmulti production.
CAUTION: addition of a field to RangeTblEntry nodes breaks stored rules;
you will have to initdb if you have any rules.
quite the same way that transformInsertStatement does, so that an expression
could be accepted by CREATE TABLE and then fail when used. Also, put back
check that CONSTRAINT expressions must yield boolean...
expressions in CREATE TABLE. There is no longer an emasculated expression
syntax for these things; it's full a_expr for constraints, and b_expr
for defaults (unfortunately the fact that NOT NULL is a part of the
column constraint syntax causes a shift/reduce conflict if you try a_expr.
Oh well --- at least parenthesized boolean expressions work now). Also,
stored expression for a column default is not pre-coerced to the column
type; we rely on transformInsertStatement to do that when the default is
actually used. This means "f1 datetime default 'now'" behaves the way
people usually expect it to.
BTW, all the support code is now there to implement ALTER TABLE ADD
CONSTRAINT and ALTER TABLE ADD COLUMN with a default value. I didn't
actually teach ALTER TABLE to call it, but it wouldn't be much work.
--enable-debug adds -g (unconditionally)
--disable-debug removes -g (if it was already in there somehow)
(giving neither does nothing)
Since none of the templates default CFLAGS with a -g you're not likely
to
end up with two -g flags. Not that they'd hurt though.
It doesn't do anything about C++.
Peter Eisentraut
not just C, so that ISCACHABLE attribute can be specified for user-defined
functions. Get rid of ParamString node type, which wasn't actually being
generated by gram.y anymore, even though define.c thought that was what
it was getting. Clean up minor bug in dfmgr.c (premature heap_close).
modifyAggrefQual. This routine really, really needs to be retired, but
until we have subselects in FROM there's no chance of doing the job right.
In the meantime try to respond to unhandlable cases with elog rather than
coredump.
in the Expr nodes they produce. This fixes a few cases of errors like
'typeidTypeRelid: Invalid type - oid = 0' caused by calling parser-related
routines on expression trees that have already been processed by planner-
related routines.
they have no hardwired limit on the length of a rule's text. Fix a couple
of minor bugs in passing --- deparsed UPDATE queries didn't have quotes
around relation name, and quotes and backslashes in constant values weren't
backslash-quoted.
expression_tree_mutator rather than ad-hoc tree walking code. This shortens
the code materially and fixes a fair number of sins of omission. Also,
change modifyAggrefQual to *not* recurse into subselects, since its mission
is satisfied if it removes aggregate functions from the top level of a
WHERE clause. This cures problems with queries of the form SELECT ...
WHERE x IN (SELECT ... HAVING something-using-an-aggregate), which would
formerly get mucked up by modifyAggrefQual. The routine is still
fundamentally broken, of course, but I don't think there's any way to get
rid of it before we implement subselects in FROM ...
Implements the CREATE CONSTRAINT TRIGGER and SET CONSTRAINTS commands.
TODO:
Generic builtin trigger procedures
Automatic execution of appropriate CREATE CONSTRAINT... at CREATE TABLE
Support of new trigger type in pg_dump
Swapping of huge # of events to disk
Jan
functions. One problem that I have encountered with the function
manager is that it does not allow the user to define type conversion
functions that convert between user types. For instance if mytype1,
mytype2, and mytype3 are three Postgresql user types, and if I wish to
define Postgresql conversion functions like
I run into problems, because the Postgresql dynamic loader would look
for a single link symbol, mytype3, for both pieces of object code. If
I just change the name of one of the Postgresql functions (to make the
symbols distinct), the automatic type conversion that Postgresql uses,
for example, when matching operators to arguments no longer finds the
type conversion function.
The solution that I propose, and have implemented in the attatched
patch extends the CREATE FUNCTION syntax as follows. In the first case
above I use the link symbol mytype2_to_mytype3 for the link object
that implements the first conversion function, and define the
Postgresql operator with the following syntax
The patch includes changes to the parser to include the altered
syntax, changes to the ProcedureStmt node in nodes/parsenodes.h,
changes to commands/define.c to handle the extra information in the AS
clause, and changes to utils/fmgr/dfmgr.c that alter the way that the
dynamic loader figures out what link symbol to use. I store the
string for the link symbol in the prosrc text attribute of the pg_proc
table which is currently unused in rows that reference dynamically
loaded
functions.
Bernie Frankpitt
an empty targetlist *and* fails to return any tuples, as will happen
for example with 'SELECT COUNT(1) FROM table WHERE ...' if the where-
clause selects no tuples. It's so nice to make a fix by diking out code,
instead of adding more...
behavior as it was, apart from forbidding minus-terminated
operators. Seems that I have to break the habit of doing before
thinking properly :-/ The point is that my second patch breaks
constructs like a & b or a ! b. This patch is to be applied
instead of any of two other today's patches.
Leon
Two patches included:
- the first one enables the use of bool variables in fields which might
become NULL.
Up to now the lib told you that NULL is not a bool variable, even if
you provide a indicator.
- the second patch checks whether a value is null and issues an error if
no indicator is provided.
Sidenote: IIRC, the variable should be left alone if the value is NULL.
ECPGlib sets it's value to 0 on NULL. Is this a violation of the
standard?
Regards
Christof
is used to find start scan position of Indexscan-s.
To speed up finding scan start position,I have changed
_bt_first() to use as many keys as possible.
I'll attach the patch here.
Regards.
Hiroshi Inoue
When drawing up a very simple "text-drawing" of how the negotiation is done,
I realised I had done this last part (fallback) in a very stupid way. Patch
#4 fixes this, and does it in a much better way.
Included is also the simple text-drawing of how the negotiation is done.
//Magnus
with no input rows, per pghackers discussions around 7/22/99. Clean up
a bunch of ugly coding while at it; remove redundant re-lookup of
aggregate info at start of each new GROUP. Arrange to pfree intermediate
values when they are pass-by-ref types, so that aggregates on pass-by-ref
types no longer eat memory. This takes care of a couple of TODO items...
Frankpitt, plus some improvements from yours truly. The simplifier depends
on the proiscachable field of pg_proc to tell it whether a function is
safe to pre-evaluate --- things like nextval() are not, for example.
Update pg_proc.h to contain reasonable cacheability information; as of
6.5.* hardly any functions were marked cacheable. I may have erred too
far in the other direction; see recent mail to pghackers for more info.
This update does not force an initdb, exactly, but you won't see much
benefit from the simplifier until you do one.
* Buffer refcount cleanup (per my "progress report" to pghackers, 9/22).
* Add links to backend PROC structs to sinval's array of per-backend info,
and use these links for routines that need to check the state of all
backends (rather than the slow, complicated search of the ShmemIndex
hashtable that was used before). Add databaseOID to PROC structs.
* Use this to implement an interlock that prevents DESTROY DATABASE of
a database containing running backends. (It's a little tricky to prevent
a concurrently-starting backend from getting in there, since the new
backend is not able to lock anything at the time it tries to look up
its database in pg_database. My solution is to recheck that the DB is
OK at the end of InitPostgres. It may not be a 100% solution, but it's
a lot better than no interlock at all...)
* In ALTER TABLE RENAME, flush buffers for the relation before doing the
rename of the physical files, to ensure we don't get failures later from
mdblindwrt().
* Update TRUNCATE patch so that it actually compiles against current
sources :-(.
You should do "make clean all" after pulling these changes.
now that sequence names are properly quoted for field defaults, mixed
case sequence names are generated. These are properly quoted in the
CREATE SEQUENCE lines, but not in the SELECT nextval lines, as per
below:
CREATE SEQUENCE "Teams_TeamID_seq" start 10 increment 1 maxvalue
2147483647 minvalue 1 cache 1 ;
SELECT nextval ('Teams_TeamID_seq');
This needs to be:
SELECT nextval ('"Teams_TeamID_seq"');
Patch included below.
--
Ross J. Reedstrom, Ph.D., <reedstrm@rice.edu>
additional argument specifying the kind of lock to acquire/release (or
'NoLock' to do no lock processing). Ensure that all relations are locked
with some appropriate lock level before being examined --- this ensures
that relevant shared-inval messages have been processed and should prevent
problems caused by concurrent VACUUM. Fix several bugs having to do with
mismatched increment/decrement of relation ref count and mismatched
heap_open/close (which amounts to the same thing). A bogus ref count on
a relation doesn't matter much *unless* a SI Inval message happens to
arrive at the wrong time, which is probably why we got away with this
sloppiness for so long. Repair missing grab of AccessExclusiveLock in
DROP TABLE, ALTER/RENAME TABLE, etc, as noted by Hiroshi.
Recommend 'make clean all' after pulling this update; I modified the
Relation struct layout slightly.
Will post further discussion to pghackers list shortly.
See attached mail for more details.
-------------------------------------------------------------------
From: "Vadim Mikheev" <vadim@krs.ru>
To: "Hiroshi Inoue" <Inoue@tpf.co.jp>
References: <000201befa94$42fe04c0$2801007e@cadzone.tpf.co.jp>
Subject: Re: elog(ERROR) in vacuum
Date: Fri, 10 Sep 1999 10:27:10 +0900
Organization: OJSC Rostelecom (Krasnoyarsk)
Message-ID: <37D85E6E.5AFA126D@krs.ru>
Hiroshi Inoue wrote:
>
> Hello Vadim,
>
> I have a question about vacuum.
>
> VACUUM has a phase like commit which calls TransactionIdCommit().
> But if elog(ERROR) occured after that,the status of transaction is
> changed from XID_COMMIT to XID_ABORT.
>
> Seems to me this causes inconsistency.
> Shoudn't AbortTransaction() be changed not to call TransacionIdAbort()
> in case of vacuum.
You're right!
As usual -:)
Vadim
Almost worked before, but forgot one place to check.
Reported by Tatsuo Ishii.
Still does not do the right thing if inserting into a non-string target
column. Should look for a type coersion later, but doesn't.
message under a kernel that only returns one packet per recv() call. This
didn't use to matter much, but it starts to get annoying with multi-megabyte
EXPLAIN VERBOSE responses...
conditions. There are some pretty bogus heuristics in prepqual.c that
try to decide whether to output CNF or DNF format; they need to be replaced,
likely. Right now the code is probably too willing to choose DNF form,
which might hurt performance in some cases that used to work OK.
But at least we have a foundation to build on.
in or_normalize, remove detection of duplicate subexpressions (since it's
highly unlikely to be worth the amount of time it takes), and introduce
a dnfify() entry point so that unintelligible backwards logic in UNION
processing can be eliminated. This is just an intermediate step ---
next thing is to look at not forcing the qual into CNF form when it would
be better off in DNF form.
This change seems necessary in conjunction with long queries, and it
cleans up some bogosity in connection with long EXPLAIN texts anyway.
Note that current libpq will accept any length error message (at least
until it runs out of memory); prior versions have a limit of 8K, but
will cleanly discard excess error text, so there shouldn't be any
big compatibility problems with old clients.
transaction abort --- before it only worked if there was exactly one level
of allocation context stacked in the blank portal. Now it does the right
thing for any depth, including zero...
was rejecting negative attnums as bogus, which of course they are not.
Add code to get_attdisbursion to produce a useful value for OID attribute,
since VACUUM does not store stats for system attributes.
Also, repair bug that's been in eqjoinsel for a long time: it was taking
the max of the two columns' disbursions, whereas it should use the min.
space consumption in pull_args, and avoid doing the full CNF transform on
operands of operator clauses, where it's really not particularly helpful.
This answers the TODO item about large numbers of OR clauses, at least
partially. I was able to do a ten-thousand-OR-clause query with about
20Mb memory consumption ... it took an obscenely long time, but it worked...
corrects flex myinput() routine so that it doesn't assume there is only
one bufferload of data. We still have the issue of getting rid of
YY_USES_REJECT so that the scanner can cope with tokens larger than its
initial buffer size.
before comparison; if fields being joined are different widths then hashing
will yield wrong answer. Also, remove hashjoinable mark from all uses of
array_eq, because array structures may have padding bytes between elements
and the pad bytes are of uncertain content. This could be revisited if
array code is cleaned up.
Modify opr_sanity regress test to complain if array_eq operator is marked
hashjoinable.
offended my aesthestic sensibility that there was so much unreadable code
doing so little. Rewritten code is about half the size, faster, and
(I hope) much more intelligible.
current transaction) are not flushed by shared-cache-inval reset message.
SI reset actually works now, for probably the first time in a long time.
I was able to run initdb and regression tests with a 16-element SI message
array, with a lot of NOTICE: cache state reset messages but no crashes.
We can't support these properly, since once the relation's physical files
are unlinked, there's no way to roll back the transaction. I suppose
we could postpone the unlink till transaction commit, but then what of
BEGIN; DROP TABLE foo; CREATE TABLE foo; ?
The code does allow dropping a table/index created in the current
transaction block, however, since the post-abort state would be that
the table doesn't exist anyway.
real name before doing lookup. We only want to index temp tables by their
real names in the relcache, to ensure there's not more than one relcache
entry for them.
has positive refcount, it is rebuilt from pg_class data. This ensures
that relcache entries will track changes made by other backends. Formerly,
a shared inval report would just be ignored if it happened to arrive while
the relcache entry was in use. Also, fix relcache to reset ref counts
to zero during transaction abort. Finally, change LockRelation() so that
it checks for shared inval reports after obtaining the lock. In this way,
once any kind of lock has been obtained on a rel, we can trust the relcache
entry to be up-to-date.
the SInval spinlock while it is calling the passed invalFunction or
resetFunction. This is necessary to avoid deadlock with lmgr change;
InvalidateSharedInvalid can be called recursively now. It should be
a good performance improvement anyway --- holding a spinlock for more
than a very short interval is a no-no.
insight that RelationFlushRelation ought to invoke smgrclose, and that the
way to make that work is to ensure that mdclose doesn't fail if the relation
is already closed (or unlinked, if we are looking at a DROP TABLE). While
I was testing that, I was able to identify several problems that we had
with multiple-segment relations. The system is now able to do initdb and
pass the regression tests with a very small segment size (I had it set to
64Kb per segment for testing). I don't believe that ever worked before.
File descriptor leaks seem to be gone too.
I have partially addressed the concerns we had about mdtruncate(), too.
On a Win32 or NFS filesystem it is not possible to unlink a file that
another backend is holding open, so what md.c now does is to truncate
unwanted files to zero length before trying to unlink them. The other
backends will be forced to close their open files by relation cache
invalidation --- but I think it would take considerable work to make
that happen before vacuum truncates the relation rather than after.
Leaving zero-length files lying around seems a usable compromise.
error/notice message lengths, and number of fields per tuple. Add
pqexpbuffer.c/.h, a frontend version of backend's stringinfo module.
This is first step in applying Mike Ansley's long-query patches,
even though he didn't do any of these particular changes...
and 1370 (timestamp(datetime)). This does not force an initdb, exactly,
but you won't see the effects of the bug fix until you do one.
BTW, OID 1358 for timespan(time) is still broken:
select timespan('21:11:26'::time);
ERROR: No such function 'time_timespan' with the specified attributes
But I couldn't figure out what it ought to be defined as, so I left it be.
references or CASE expressions, didn't parenthesize complex expressions
properly. Also, always output variable references as fully qualified
names to eliminate ambiguity bug recently reported. (This could be
smarter, but reliability comes first.)
Most parts of the planner should ignore, or indeed never even see, uplevel
Vars because they will be or have been replaced by Params. There were a
couple of places that got it wrong though, probably my fault from recent
changes...
last loop which would return the *first* surviving-to-that-point candidate
regardless of which one actually passed the test. This was producing
such curious results as 'oid % 2' getting translated to 'int2(oid) % 2'.
documented intepretation of the lefthand and oper fields. Fix a number of
obscure problems while at it --- for example, the old code failed if the parser
decided to insert a type-coercion function just below the operator of a
SubLink.
CAUTION: this will break stored rules that contain subplans. You may
need to initdb.
It will keep track the number of pages allocated so that
vacuum could allocate twice of the previous allocation.
This will greatly reduce the total memory consumption of
vacuum.
ALLOC_BIGCHUNK_LIMIT are always allocated as separate malloc() blocks,
and are free()d immediately upon pfree(). Also, if such a chunk is enlarged
with repalloc(), translate the operation into a realloc() so as to
minimize memory usage. Of course, these large chunks still get freed
automatically if the alloc set is reset.
I have set ALLOC_BIGCHUNK_LIMIT at 64K for now, but perhaps another
size would be better?
match then it tried for a self-commutative operator with the reversed input
data types. This is pretty silly; there could never be such an operator,
except maybe in binary-compatible-type scenarios, and we have oper_inexact
for that. Besides which, the oprsanity regress test would complain about
such an operator. Remove nonfunctional code and simplify routine calling
convention accordingly.
and fix_opids processing to a single recursive pass over the plan tree
executed at the very tail end of planning, rather than haphazardly here
and there at different places. Now that tlist Vars do not get modified
until the very end, it's possible to get rid of the klugy var_equal and
match_varid partial-matching routines, and just use plain equal()
throughout the optimizer. This is a step towards allowing merge and
hash joins to be done on expressions instead of only Vars ...
sort order down into planner, instead of handling it only at the very top
level of the planner. This fixes many things. An explicit sort is now
avoided if there is a cheaper alternative (typically an indexscan) not
only for ORDER BY, but also for the internal sort of GROUP BY. It works
even when there is no other reason (such as a WHERE condition) to consider
the indexscan. It works for indexes on functions. It works for indexes
on functions, backwards. It's just so cool...
CAUTION: I have changed the representation of SortClause nodes, therefore
THIS UPDATE BREAKS STORED RULES. You will need to initdb.
above a Sort or Materialize node. As far as I can tell, the only place
that actually needed that was set_tlist_references, which was being lazy
about checking to see if it had a noname node to fix or not...
store all ordering information in pathkeys lists (which are now lists of
lists of PathKeyItem nodes, not just lists of lists of vars). This was
a big win --- the code is smaller and IMHO more understandable than it
was, even though it handles more cases. I believe the node changes will
not force an initdb for anyone; planner nodes don't show up in stored
rules.
commuted (ie, the index var appears on the right). These are now handled
the same way as merge and hash join quals that need to be commuted: the
actual reversing of the clause only happens if we actually choose the path
and generate a plan from it. Furthermore, the clause is only reversed in
the 'indexqual' field of the plan, not in the 'indxqualorig' field. This
allows the clause to still be recognized and removed from qpquals of upper
level join plans. Also, simplify and generalize match_clause_to_indexkey;
now it recognizes binary-compatible indexes for join as well as restriction
clauses.
contains much code that looks like it will handle indexquals with the index
key on either side of the operator, in fact indexquals must have the index
key on the left because of limitations of the ScanKey machinery. Perhaps
someone will be motivated to fix that someday...
work under a wider range of scenarios than it did --- it formerly did not
handle a multi-pass inner scan, nor cases in which the inner scan's
indxqualorig or non-index qual contained outer var references. I am not
sure that these limitations could be hit in the existing optimizer, but
they need to be fixed for future expansion.
> >
> > was implemented by Jan Wieck.
> > His work is for ascending order cases.
> >
> > Here is a patch to prevent sorting also in descending
> > order cases.
> > Because I had already changed _bt_first() to position
> > backward correctly before v6.5,this patch would work.
> >
Hiroshi Inoue
Inoue@tpf.co.jp
multi-scan indexscan plans; it tried to use the same table-to-index
attribute mapping for all the scans, even if they used different indexes.
It would klugily work as long as OR indexquals never used multikey indexes,
but that's not likely to hold up much longer...
to go along with expression_tree_walker. (_walker is not suitable for
routines that need to alter the tree structure significantly.) Other minor
cleanups in clauses.c.
Also, move responsibility for calling vc_abort into main xact.c list of
things-to-call-at-abort. What in the world was it doing down inside of
TransactionIdAbort()?
hashjoinable clause, not one path for a randomly-chosen element of each
set of clauses with the same join operator. That is, if you wrote
SELECT ... WHERE t1.f1 = t2.f2 and t1.f3 = t2.f4,
and both '=' ops were the same opcode (say, all four fields are int4),
then the system would either consider hashing on f1=f2 or on f3=f4,
but it would *not* consider both possibilities. Boo hiss.
Also, revise estimation of hashjoin costs to include a penalty when the
inner join var has a high disbursion --- ie, the most common value is
pretty common. This tends to lead to badly skewed hash bucket occupancy
and way more comparisons than you'd expect on average.
I imagine that the cost calculation still needs tweaking, but at least
it generates a more reasonable plan than before on George Young's example.
(it should just call the given operator, not look up an = operator).
Fix intltsel() so that all numeric data types are converted to double
before trying to estimate where the given comparison value is in the
known range of column values. intltsel() still needs work, or replacement,
for non-numeric data types ... but for nonintegral numeric types it
should now be delivering reasonable estimates.
configure.in to determine if a system is ELF or not. Note that some
of the tests earlier may be redundant but I took the safest route.
D'Arcy J.M. Cain
neqsel now behave as per my suggestions in pghackers a few days ago.
selectivity for < > <= >= should work OK for integral types as well, but
still need work for nonintegral types. Since these routines have never
actually executed before :-(, this may result in some significant changes
in the optimizer's choices of execution plans. Let me know if you see
any serious misbehavior.
CAUTION: THESE CHANGES REQUIRE INITDB. pg_statistic table has changed.
so that Case works in WHERE join clauses. Temporary patch --- this routine
is one of many that ought to be changed to use centralized expression-tree-
walking logic.
rels that the inner path needs to join to, but it was only checking for
the first one. Failure could only have been observed with an OR-clause
that mentions 3 or more tables, and then only if the bogus path was
actually selected as cheapest ...
optimizer rather than parser. This has many advantages, such as not
getting fooled by chance uses of operator names ~ and ~~ (the operators
are identified by OID now), and not creating useless comparison operations
in contexts where the comparisons will not actually be used as indexquals.
The new code also recognizes exact-match LIKE and regex patterns, and
produces an = indexqual instead of >= and <=.
This change does NOT fix the problem with non-ASCII locales: the code
still doesn't know how to generate an upper bound indexqual for non-ASCII
collation order. But it's no worse than before, just the same deficiency
in a different place...
Also, dike out loc_restrictinfo fields in Plan nodes. These were doing
nothing useful in the absence of 'expensive functions' optimization,
and they took a considerable amount of processing to fill in.
The only place it was being used was as temporary storage in indxpath.c,
and the logic was wrong: the same restrictinfo node could get chosen to
carry the info for two different joins. Right fix is to return a second
list of unjoined-relids parallel to the list of clause groups.
identified by Hiroshi (incorrect cost attributed to OR clauses
after multiple passes through set_rest_selec()). I think the code
was trying to allow selectivities of OR subclauses to be passed in
from outside, but noplace was actually passing any useful data, and
set_rest_selec() was passing wrong data.
Restructure representation of "indexqual" in IndexPath nodes so that
it is the same as for indxqual in completed IndexScan nodes: namely,
a toplevel list with an entry for each pass of the index scan, having
sublists that are implicitly-ANDed index qual conditions for that pass.
You don't want to know what the old representation was :-(
Improve documentation of OR-clause indexscan functions.
Remove useless 'notclause' field from RestrictInfo nodes. (This might
force an initdb for anyone who has stored rules containing RestrictInfos,
but I do not think that RestrictInfo ever appears in completed plans.)
the query string to handle any length, I discovered that under certain
conditions, psql will core dump when handling long strings. Thus, the
patch. It was caused by a buffer overrun, probably not noticeable in a lot
of cases, but pretty noticeable in mine.
Problem was caused by the fact that the length check is only performed after
the check for a ; to get the end of the query and execute.
Cheers...
MikeA
support, but which the grammar was accepting. Also, fix several bugs
having to do with failure to copy fields up from a subselect to a select
or insert node.
of the SELECT part of the statement is just like a plain SELECT. All
INSERT-specific processing happens after the SELECT parsing is done.
This eliminates many problems, e.g. INSERT ... SELECT ... GROUP BY using
the wrong column labels. Ensure that DEFAULT clauses are coerced to
the target column type, whether or not stored clause produces the right
type. Substantial cleanup of parser's array support.
creates a reduce/reduce conflict, which I resolved by changing the
'AexprConst -> Typename Sconst' rule to 'AexprConst -> SimpleTypename Sconst'.
In other words, a subscripted type declaration can't be used in that
syntax any longer. This seems a small price to pay for not having to
qualify subscripted columns anymore.
Other cleanups: rename res_target_list to update_target_list, and remove
productions for variants that are not legal in an UPDATE target list;
rename res_target_list2 to plain target_list; delete position_expr
in favor of using b_expr in that production; merge opt_indirection
into attr nonterminal, since there are no places where an unsubscripted
attr is wanted; fix typos in Param support; change case_arg so that
an arbitrary a_expr is allowed, not only a column name.
care of equal-key cases, eliminating bt_firsteq(). The linear search
formerly done by bt_firsteq() took a lot of time in the case where many
equal keys appear on the same page.
that contain null fields. Old code would produce erratic sort results
because comparisons of tuples containing nulls could produce inconsistent
answers.
> the DTK_MICROSEC case is just like the DTK_MILLISEC case.
> I think this is wrong and it ought to look like
> fsec = rint(fsec * 1000000) / 1000000;
> no?
Tom Lane.
"HAS_LONG_LONG" is defined based on the assumption that
strtol() would return ERANGE if a platform does not support
64-bit integers. In current PostgreSQL 6.5 (and 6.4.2)
distribution, "HAS_LONG_LONG" is defined only if platform
is "alpha". (See include/port/alpha.h) I think the int4
range check should apply to linux_alpha as well. (I have
not tested yet but I guess this might be applicable to
newer Linux/i386 distributions which includes new GCC which
implements long int as 64-bit int.)
with expression_tree_walker-based code. The former failed to cope with
expressions containing SubLinks, and the latter returned TRUE for both
SubLinks and Aggrefs (cut-and-paste bug?). There is a lot more scope for
using expression_tree_walker in this module, but I'll restrain myself
until the 6.6 split occurs from touching not-demonstrably-broken code.
is parse_aggs.c. This fixes its failure to cope with (at least) CaseExpr
and ArrayRef nodes, which is the reason why both of these fail in 6.5:
select coalesce(f1,0) from int4_tbl group by f1;
ERROR: Illegal use of aggregates or non-group column in target list
select sentence.words[0] from sentence group by sentence.words[0];
ERROR: Illegal use of aggregates or non-group column in target list
The array case still fails, but at least it's not parse_agg's fault
anymore ... considering that we now support CASE officially, I think
it's important to fix the first example ...
will gradually replace all of the boilerplate tree-walk-recursion code that
currently exists in O(N) slightly different forms in N subroutines.
I've had it with adding missing cases to these subroutines...
August 1994 draft standard.
Use the ecpg support libraries to write the CLI interface?
Date and Darwen claim that CLI is a more modern and flexible approach...
special hack to ensure it would close its output file even after failure
due to elog(ERROR) partway through the copy. This is now unnecessary
because fd.c takes care of cleaning up open files at transaction abort;
worse, after fd.c closed the file copy.c would try to do so *again* at
the start of the next COPY command. This would result in havoc in most
implementations of stdio library.
tlist and qual are NULL. It ought to handle these the same as the cases
where tlist contains only constant expressions, ie, be willing to generate
a Result-node plan. This did not use to matter, but it does now because
union_planner will flatten the tlist when aggregates are present. Thus,
'select count(1) from table' now causes query_planner to be given a null
tlist, and to duplicate 6.4's behavior we need it to give back a Result
plan rather than refusing the query. 6.4 was arguably doing the Wrong
Thing for this query, but I'm not going to open a semantics issue right
before 6.5 release ... can revisit that problem later.
returned NULL, which it will do in some cases where an elog(ERROR) would
probably be more appropriate. For the moment, generate a not-very-
informative error message rather than proceeding to certain coredump.
Probably ought to think about making query_planner elog instead of
returning NULL, but this is at least a safe change for now.
pointer to palloc'd but uninitialized memory. This is not cool; anyone looking
at the returned 'tuple' would at best coredump and at worst behave in a
bizarre and irreproducible way. Fix it to return a predictable value,
namely a correctly-set-up palloc'd tuple containing zero attributes.
I believe this fix is both safe and critical.
this one could be useful for people experiencing out-of-memory crashes while
executing queries which retrieve or use a very large number of tuples.
The problem happens when storage is allocated for functions results used in
a large query, for example:
select upper(name) from big_table;
select big_table.array[1] from big_table;
select count(upper(name)) from big_table;
This patch is a dirty hack that fixes the out-of-memory problem for the most
common cases, like the above ones. It is not the final solution for the
problem but it can work for some people, so I'm posting it.
The patch should be safe because all changes are under #ifdef. Furthermore
the feature can be enabled or disabled at runtime by the `free_tuple_memory'
options in the pg_options file. The option is disabled by default and must
be explicitly enabled at runtime to have any effect.
To enable the patch add the follwing line to Makefile.custom:
CUSTOM_COPT += -DFREE_TUPLE_MEMORY
To enable the option at runtime add the following line to pg_option:
free_tuple_memory=1
Massimo
/*
* Read above about cases when !ItemIdIsUsed(Citemid)
* (child item is removed)... Due to the fact that
* at the moment we don't remove unuseful part of
* update-chain, it's possible to get too old
* parent row here. Like as in the case which
* caused this problem, we stop shrinking here.
* I could try to find real parent row but want
* not to do it because of real solution will
* be implemented anyway, latter, and we are too
* close to 6.5 release. - vadim 06/11/99
*/
if (Ptp.t_data->t_xmax != tp.t_data->t_xmin)
...
1. Using 100 digits after decimal point on the default
make runtest.
2. Using 1000 digits after decimal point in a new target
make bigtest.
At the end of 'make runtest', a hint about the new bigtest is
printed.
Jan
and possibly for other cases too:
DO NOT cache status of transaction in unknown state
(i.e. non-committed and non-aborted ones)
Example:
T1 reads row updated/inserted by running T2 and cache T2 status.
T2 commits.
Now T1 reads a row updated by T2 and with HEAP_XMAX_COMMITTED
in t_infomask (so cached T2 status is not changed).
Now T1 EvalPlanQual gets updated row version without HEAP_XMIN_COMMITTED
-> TransactionIdDidCommit(t_xmin) and TransactionIdDidAbort(t_xmin)
return FALSE and T2 decides that t_xmin is not committed and gets
ERROR above.
It's too late to find more smart way to handle such cases and so
I just changed xact status caching and got rid TransactionIdFlushCache()
from code.
Changed: transam.c, xact.c, lmgr.c and transam.h - last three
just because of TransactionIdFlushCache() is removed.
2. heapam.c:
T1 marked a row for update. T2 waits for T1 commit/abort.
T1 commits. T3 updates the row before T2 locks row page.
Now T2 sees that new row t_xmax is different from xact id (T1)
T2 was waiting for. Old code did Assert here. New one goes to
HeapTupleSatisfiesUpdate. Obvious changes too.
3. Added Assert to vacuum.c
4. bufmgr.c: break
Assert(buf->r_locks == 0 && !buf->ri_lock)
into two Asserts.
after ExecEndNode. It must be done! Or we'll be out of free
tuple slots very soon, though slots are freed by ExecEndNode
and ready for reusing.
We didn't see this problem before because of
int nSlots = ExecCountSlotsNode(plan);
TupleTable tupleTable = ExecCreateTupleTable(nSlots + 10);
/* why add ten? - jolly */
code in InitPlan - i.e. extra 10 slots. Simple select uses
3 slots and so it was possible to re-use evaluation plan
3 additional times and didn't get
elog(NOTICE, "Plan requires more slots than are available");
elog(ERROR, "send mail to your local executor guru to fix this");
Changes are obvious and shouldn't be problems with them.
Though, I added Assert(epqstate->es_tupleTable->next == 0)
before EvalPlanQual():ExecInitNode and we'll notice if
something is still wrong. Is it better to change Assert
to elog(ERROR) ?
1. check whether the program is being executed in $PGDATA/.. This is
necessary if the data tree is not in the standard place, as is the
case with the Debian distribution (because of Debian policy).
2. give a clearer error message if the dumped data structure fails to
be loaded.
Oliver Elphick
SHARED_LIB:
needs to be changed to:
SHARED_LIB:-lc
I think this was also needed on AIX 4.2. Comments Please !!
If nobody objects, I suggest to make this change, since it cannot
break AIX 4.2 and is necessary on AIX 4.3
Andreas
> (native win32, not cygnus).
> It does the following:
> Patches two win32.mak files to DEFINE HAVE_VSNPRINTF and
> HAVE_STRDUP. This is required to build at all.
> Bumps the version number on libpq.dll from 6.4 to 6.5.
> Required for install programs to work.
> Adds defintions for BLCKSZ and MAXIMUM_ALIGN to "win32.h" in
> the client-side libpiq directory.
>
> All these files are only used when building on native win32,
> so it should be safe I think.
>
> Again, really sorry to throw this in so late, but I would
> hate to do the same thing as with 6.4 (which required 6.4.1
> to at all compile on Win32).
>
> Thanks,
>
> //Magnus
a non-leading % would be put into the >=/<= patterns. Also, repair
longstanding confusion about whether %% means a literal %%. The SQL92
doesn't say any such thing, and textlike() knows that, but gram.y didn't.
2. varsup.c:ReadNewTransactionId(): don't read nextXid from disk -
this func doesn't allocate next xid, so ShmemVariableCache->nextXid
may be used (but GetNewTransactionId() must be called first).
3. vacuum.c: change elog(ERROR, "Child item....") to elog(NOTICE) -
this is not ERROR, proper handling is just not implemented, yet.
4. s_lock.c: increase S_MAX_BUSY by 2 times.
5. shmem.c:GetSnapshotData(): have to call ReadNewTransactionId()
_after_ SpinAcquire(ShmemIndexLock).
PgDatabase::DisplayTuples and PgDatabase::PrintTuples. This is incorrect
according to strict interpretation of the C++ spec, and some compilers
will reject it. Also silence g++ warning about unused parameter.
the gettimeofday doesn't compile under Linux with glibc2 because
the DST_NONE constant is no more defined. It seems that this code
(written by me) has always be wrong but for some reason working.
From: Massimo Dal Zotto <dz@cs.unitn.it>
the default target is 'install' instead of 'all'. So if you do a
make without target you actually do a make install, which is not
what one normally expects from a standard makefile.
From: Massimo Dal Zotto <dz@cs.unitn.it>
the ecpg Makefiles use a variable DESTDIR which is never defined
except by debian/rules makefile, in which case the ecpg makefiles
expand wrong pathnames. If we want to support a DESTDIR root it
must be done consistently in all the makefiles, not just in ecpg.
From: Massimo Dal Zotto <dz@cs.unitn.it>
they were confusing because the large object tables themselves are not
shown. (Besides, if you've got hundreds or thousands of large objects,
you really don't want to see 'em at all.)
Also, suppress all indexes from the \z ACL listing, since indexes have
no meaningful protection information.
SET TRANSACTION ISOLATION LEVEL READ COMMITTED;
^^^^ required
Also note that SET TRANSACTION ISOLATION LEVEL change
isolevel for _current_ transaction, in accordance with
standard, not for session (ALTER SESSION is used in Oracle,
but it's not implemented currently).
And I would don't mention SET XACTISOLEVEL TO ...
form at all.
Please update set.sgml - I failed to understand all these
SET TIME ZONE { '<REPLACEABLE CLASS="PARAMETER">
now.
for Vadim
transactions will not assume that MyProc transaction was committed
before snapshot calculations. With old MyProc->xid assignment
(in xact.c:StartTransaction()) there was ability to see the same
row twice (I used gdb for this)!...
2. Assignments of InvalidTransactionId to MyProc->xid and MyProc->xmin
are moved from xact.c:CommitTransaction() to
xact.c:RecordTransactionCommit() - this invalidation must be done
before releasing transaction locks or bad (too high) XmaxRecent value
might be used by vacuum ("ERROR: Child itemid marked as unused"
reported by "Hiroshi Inoue" <Inoue@tpf.co.jp>; once again, gdb
allowed me reproduce this error).
{
Oid relId;
Oid dbId;
union
{
BlockNumber blkno;
TransactionId xid;
} objId;
>
> Added:
> /*
> * offnum should be part of objId.tupleId above, but would increase
> * sizeof(LOCKTAG) and so moved here; currently used by userlocks only.
> */
> OffsetNumber offnum;
uint16 lockmethod; /* needed by userlocks */
} LOCKTAG;
gmake clean required...
User locks are ready for 6.5 release...
2. Get rid of locking when updating statistics in vacuum.
3. Use QuerySnapshot in COPY TO and call SetQuerySnashot
in main tcop loop before FETCH and COPY TO.
redundant) SearchSysCache searches per table column in an INSERT, which
accounted for a good percentage of the CPU time for INSERT ... VALUES().
Now it only does two searches in the typical case.
through MAXBACKENDS array entries used to be fine when MAXBACKENDS = 64.
It's not so cool with MAXBACKENDS = 1024 (or more!), especially not in a
frequently-used routine like SIDelExpiredDataEntries. Repair by making
procState array size be the soft MaxBackends limit rather than the hard
limit, and by converting SIGetProcStateLimit() to a macro.
do the right thing: look for a NOTICE message from the backend before we
close our side of the socket. 6.4 libpq did not reliably print the backend's
hara-kiri message, 'The Postmaster has informed me ...', because it only
did the right thing if connection closure was detected during a read
attempt instead of a write attempt.
BT_READ/BT_WRITE are BUFFER_LOCK_SHARE/BUFFER_LOCK_EXCLUSIVE now.
Also get rid of #define BT_VERSION_1 - we use version 1 as default
for near two years now.
not be marked inFromCl any longer. Otherwise the planner gets confused
and joins over them where in fact it does not have to.
Adjust hasSubLinks now with a recursive lookup - could be wrong in
multi action rules because parse state isn't reset correctly and all
actions in the rule are marked hasSubLinks if one of them has.
Jan
aggregate functions, as in
select a, b from foo group by a;
The ungrouped reference to b is not kosher, but formerly we neglected to
check this unless there was an aggregate function somewhere in the query.
this file in interfaces/
It will all need to be checked in. I used the char *rcsid[] method for
cvs ids so it can be strings | grep'd to find version numbers. The new
version for the library is 3.0.
Run configure from src/ to create the Makefile and it should be good to
go.
I did minimal documentation references in the README, I'll see if I can
get something to Tom Lockhart rather quickly.
Vince.
SelectStmt and CursorStmt tried to parse FOR UPDATE ... / FOR READ ONLY.
Cursor now checks that it is read only by looking at forUpdate of Query.
SelectStmt handles FOR READ ONLY too.
Jan
will pass through rather than spitting up. This is necessary to handle
cases where coerce_type causes a subexpression to be retransformed, as in
SELECT count(*) + 1.0 FROM table
remove optimizer's arbitrary limit on how large a join it will use hashing
for. (The limit was too large to prevent the problems we'd been seeing,
anyway...)
fixed-size hashtable. This should prevent 'hashtable out of memory' errors,
unless you really do run out of memory. Note: target size for hashtable
is now taken from -S postmaster switch, not -B, since it is local memory
in the backend rather than shared memory.
looks
like someone just didn't add support for multiple segments for
truncation.
The following patch seems to do the right thing, for me at least.
It passed my tests, my data looks right(no data that shouldn't be in
there) and regression is ok.
Ole Gjerde
segments, and my indexes had 3(Yes, it DOES work!).
DROP TABLE removed ALL segments from the table, but only the main index
segment.
So it looks like removing the table itself is using mdunlink in md.c,
while removing indexes uses FileNameUnlink() which only unlinks 1 file.
As far as I can tell, calling FileNameUnlink() and mdunlink() is basically
the same, except mdunlink() deletes any extra segments.
I've done some testing and it seems to work. It also passes regression
tests(except float8, geometry and rules, but that's normal).
If this patch is right, this fixes all known multi-segment problems on
Linux.
Ole Gjerde
configtype.patch simply fixes a typo in config.h.in
pg_dump.c.patch Updates a bunch of error messages to include a reason
from
the backend, and also removes a couple of unnecessary
if's
Ole Gjerde
lists are now plain old garden-variety Lists, allocated with palloc,
rather than specialized expansible-array data allocated with malloc.
This substantially simplifies their handling and eliminates several
sources of memory leakage.
Several basic types of erroneous queries (syntax error, attempt to
insert a duplicate key into a unique index) now demonstrably leak
zero bytes per query.
The
offending code
has been removed, the action is now always dependent :-)
I suggest the following patch, to finally make trigger regression happy
again:
<<refint1.patch>>
After that you can remove the following from TODO:
Remove ERROR: check_primary_key: even number of arguments should be
specified
Trigger regression test fails
Andreas
and lock syntax as fully parsed tokens.
Two keywords for isolation are non-reserved SQL92
(COMMITTED, SERIALIZABLE).
All other new keywords are non-reserved Postgres (not SQL92)
(ACCESS, EXCLUSIVE, MODE, SHARE).
Add syntax to allow CREATE [GLOBAL|LOCAL] TEMPORARY TABLE, throwing an
error if GLOBAL is specified.
constraints. Reported by Tom Lane.
Now, check for duplicate indices and retain the one which is a primary-key.
Adjust elog NOTICE messages to surround table and column names with single
quotes.
-d4 now prints compressed trees from nodeToString()
-d5 prints pretty trees via nodeDisplay()
new pg_options: pretty_plan, pretty_parse, pretty_rewritten
Jan
on connection. This patch changes it to use PQconnectdb rather than
{fe_setauthsvc,PQsetdb}. This still isn't the complete solution, as
there
is no provision for user,password in class PgEnv, but it does get rid of
the error message. Tested with gcc version egcs-2.91.60 19981201
(egcs-1.1.1 release) under NetBSD-1.3K/i386.
Cheers,
Patrick Welche
files to be closed automatically at transaction abort or commit, should
they still be open. Also close any still-open stdio files allocated with
AllocateFile at abort/commit. This should eliminate problems with leakage
of file descriptors after an error. Also, put in some primitive buffered-IO
support so that psort.c can use virtual files without severe performance
penalties.
"SYSTEM", and unpack the files in the uuencoded .tar.gz file at the end in
src/test/regress so that the int2, int4 and geometry tests pass on NetBSD/i386.
They just fail on different wording of error messages and eg printing "0"
rather than "-0". At a guess the same will be true for the other NetBSD ports,
but I can't test them.
Cheers,
Patrick
about certain to fail anytime it decided the relation to be hashed was
too big to fit in memory --- the code for 'batching' a series of hashjoins
had multiple errors. I've fixed the easier problems. A remaining big
problem is that you can get 'hashtable out of memory' if the code's
guesstimate about how much overflow space it will need turns out wrong.
That will require much more extensive revisions to fix, so I'm committing
these fixes now before I start on that problem.
arrayfuncs.patch fixes a small bug in my previous patches for
arrays
array-regress.patch adds _bpchar and _varchar to regression tests
--
Massimo Dal Zotto
been applied. The patches are in the .tar.gz attachment at the end:
varchar-array.patch this patch adds support for arrays of bpchar() and
varchar(), which where always missing from postgres.
These datatypes can be used to replace the _char4,
_char8, etc., which were dropped some time ago.
block-size.patch this patch fixes many errors in the parser and other
program which happen with very large query statements
(> 8K) when using a page size larger than 8192.
This patch is needed if you want to submit queries
larger than 8K. Postgres supports tuples up to 32K
but you can't insert them because you can't submit
queries larger than 8K. My patch fixes this problem.
The patch also replaces all the occurrences of `8192'
and `1<<13' in the sources with the proper constants
defined in include files. You should now never find
8192 hardwired in C code, just to make code clearer.
--
Massimo Dal Zotto
to save a little bit of backend startup time. This way, the first
backend started after a VACUUM will rebuild the init file with up-to-date
statistics for the critical system indexes.
the database encoding and the client encoding match the encoding expected
by the test. So, force both of them to be set from the MULTIBYTE
environment var. This allows regress tests to be run successfully in
multibyte environments other than the compiled-in default.
it failed to cover the case where high bits of char are 100 or 101.
Not sure if fix is right, but it agrees with pg_utf_mblen ... and it
doesn't lock up ...
can be generated in a buffer and then sent to the frontend in a single
libpq call. This solves problems with NOTICE and ERROR messages generated
in the middle of a data message or COPY OUT operation.
instead of doing a kill(self, SIGQUIT) and expecting the signal handler
to do it. Also, clean up inconsistent definitions of the sigjmp buffer
in the several files that already referenced it.
and related files. Also remove float.c's gratuitous redeclaration of
isinf() ... looks like there are more decls in there that ought to be
in config.h, but I'll leave well enough alone for now ...
expression context (ie, not at the top level of a WHERE clause). Examples
like this one work now:
SELECT name, value FROM t1 as touter WHERE
(value/(SELECT AVG(value) FROM t1 WHERE name = touter.name)) > 0.75;
delete the default argument from the node. This prevents the executor
from spitting up on the untransformed argument expression. Typical
failure was:
select (case f1 when 'val' then 'subst' else f1 end) from t1;
ERROR: copyObject: don't know how to copy 704
MyProcPid global variable is set to 0 when postgres starts as a command
(not as a backend daemon). This leads issuing SIGQUIT to the process group,
not the process itself. As a result, parent sh gets core dumped in the
Wisconsin benchmark test.
- change temp -> temp_bench ("temp" is now a reserved word)
- fix bugs in queries
- add -B 256 option to run the postgres command
(without this, postgres seems to fail with hashjoin)
rather than reusing the input storage.
Also made the same fix to int8smaller(), though there wasn't a symptom,
and went through and verified that other pass-by-reference data types
do the same thing. Not an issue for the by-value types.
relation, rather than zeroes. This prevents the optimizer from making
foolish choices (ie, using nested-loop plans) on never-yet-vacuumed tables.
This is a hack, of course. Keeping accurate track of these statistics
would be a cleaner solution, but it's far from clear that it'd be worth
the cost of doing so. In any case we're not going to do that for 6.5.
In the meantime, this quick hack provides a useful performance improvement
in the regression tests and in many real-world scenarios.
in rules regression test, in order to eliminate bogus test 'failures'
that occur due to platform-dependent and join-implementation-dependent
ordering of tuples. I'm not sure that I got all of the SELECTs that need
ordering clauses --- we may need some more. But this takes care of the
diffs between my platform and Jan's.
sourced with \i (tried to read data from the terminal, rather than from
the source file; this breaks pg_dump scripts read with \i). Also, \o file
followed by COPY TO STDOUT wrote to terminal not designated file.
All better now.
time zone.
Previously, localtime() rotated a date with a day of month field which
exceeded the actual range into the next months, masking the fact that
a bad date had been specified.
Regression tests pass.
Previously, dates falling within Unix system time range were run through
a call to localtime() to get the time zone, if it was not specified.
This had the effect that dates with DOMs which were larger than would be
valid for that month were "rotated" into the following months.
syntax for CREATE OPERATOR with SORT parameters. Fixed.
It is now actually possible to dump and reload a database containing
fully specified user-definable operators ...
indexes.
1. Index Scan using plural indexids never scan backward
as to the order of indexids.
2. The cursor using Index scan is not usable after moving
past the end.
This patch solves above bugs.
Moreover the change of _bt_first() would be useful to extend
ORDER BY patch by Jan Wieck for all descending order cases.
Hiroshi Inoue
not-yet-defined operator in commutator, negator, etc links. This is
necessary in order to ensure that a pg_dump dump of user-defined operators
can be reloaded. There may still be a bug lurking here, because it's
provoking a 'Buffer Leak' notice message in one case. See my mail to
pgsql-hackers.
hashjoin's hashFunc() so that it does the right thing with pass-by-value
data types (the old code would always return 0 for int2 or char values,
which would work but would slow things down a lot). Extend opr_sanity
regress test to catch more kinds of errors.
called through fmgr. Someday we should try to actually execute the function,
but that looks like it might be a major feature addition.
Not something to try during beta phase.
it with configure-script tests to see whether const, inline, volatile, etc
work or not. (Curiously, configure was already doing the work to see if
const and inline were OK, but the results were not getting plugged into
config.h :-(.)
1. Fix problems of PAGER and \? command
2. Add -E option that shows actual queries sent by \dt and friends
3. Add version number in startup banners for psql
There are two subdirectories (ISO8859-7 and koi8-to-win1251) containing
tests for Greek locale and server<=>client recoding feature (recently
submitted by Tatsuo Ishii <t-ishii@sra.co.jp>; we've debugged his patches
together in the field of Cyrillic support).
function is found in prosrc field of pg_proc, not proname. This allows
multiple aliases of a built-in to all be implemented as direct builtins,
without needing a level of indirection through an SQL function. Replace
existing SQL alias functions with builtin entries accordingly.
Save a few K by not storing string names of builtin functions in fmgr's
internal table (if you really want 'em, get 'em from pg_proc...).
Update opr_sanity with a few more cross-checks.
2. Much faster btree tuples deletion in the case when first on page
index tuple is deleted (no movement to the left page(s)).
3. Remember blkno of new root page in BTPageOpaque of
left/right siblings when root page is splitted.
I have solved some problems with dynamic loading on NT. It is possible
to
run succesfully both trigger and plpgsql regression tests. The patch is
in
the included file "diff".
Dan
but others declare it as extern char *. gcc complains (quite rightly too).
Worked around it by rearranging the order of inclusions so that we don't
have to explicitly declare yytext; this should work with either variant.
change functionality, but makes the code more ANSI C'ish.
My AIX xlc compiler barfs on all of these. Can someone please
review and apply to current.
<<port.patch>>
Thanks
Andreas
in
different directories. The patch gave the option of specifying a dir
for
the tk script and if they were both in the same directory then it didn't
mind being empty. It's small so I'm including it. It was tested with
autoconf 2.12.
Vince.
would be a Bad Thing.
For what it's worth, I found another case in libpq where you can get a T
message without a D that my utility patch needs to handle. I have
attached
the updated patch against the 6.4.2 version of
src/interfaces/libpq/fe-exec.c
Jerry Gay
results in a bogus datetime value under AlphaLinux. (Note that
the link to submit a port-specific bug on your website is broken)
-Test Case:
----------
testdb=> create table dttest (dt datetime);
testdb=> insert into dttest values ('now');
--------------------------------------------------------------------------
Solution:
---------
The basic problem is the typedefs of AbsoluteTime and RelativeTime,
which are both 'int32'. These types appear to be used synonymously
with the 'time_t' type, which on AlphaLinux is typedef'd as a 'long
int', which is 64-bits (not 32). The solution included here fixes
the datetime type (it now passes the regression test), but does not
pass the absolute and relative time regression tests. Presumably, a
more thorough investigation of how these types are used is warranted.
The included patch is from the v6.3.2 source, but can be applied to
the v6.4.2 source. Please note that there is also a RedHat-specific
patch distributed with the PostgreSQL source package from RedHat
that was applied first.
Rich Edwards
Just in case you'd like to see what I was talking about, I am
attaching
my patch to src/interfaces/libpq/fe-exec.c to prevent utility functions
called from SPI from locking up the client.
Jerry Gay
authentifica
tion
working with postgresql-6.4.2 and KTH-KRB Ebones
(http://www.pdc.kth.se/kth-kr
b) on a dec alpha running DU 4.0D using the native compiler. The
following
patch does the trick.
The rationale behind this is as follows. The KTH-KRB code header files
defines
lots of lengths like INST_SZ,REALM_SZ and KRB_SENDAUTH_VLEN. It also has
a
habit of doing things like
chararray[LENGTH] = '\0'
to ensure null terminated strings. In my instance this just happens to
blat
the kerberos principal instance string leading to error like
pg_krb4_recvauth: kerberos error: Can't decode authenticator
(krb_rd_req
)
The application code that comes with KTH-KRB uses "KRB_SENDAUTH_VLEN +
1" and
sometimes uses "INST_SZ + 1" so it seems safest to put that 1 char
buffer in
the appropriate place.
Rodney McDuff
NetBSD/macppc
LinuxPPC
FreeBSD 2.2.6-RELEASE
All of them seem happy with the regression test. Note that, however,
compiling with optimization enabled on NetBSD/macppc causes an initdb
failure (other two platforms are ok). After checking the asm code, we
are suspecting that might be a compiler(egcs) bug.
Tatsuo Ishii
Removed char16 and replaced with an example using Point
as suggested by Tom Lane. The dept field was changed to
the cubicle field denoting the row(x) and column(y) of
the employee's cube in the corporate jungle. The C function
builds a 'compromise' cubicle from two suggested ones.
I'll try and patchup the documentation next.
Clark
I've been working on the following TODO list item:
* psql \d on index with char()/varchar() fields shows improper length
I've attached a simple patch to fix this.
-Ryan
records using a sub form, i.e. entering a new order/orderlines or master
and
detail tables. The problem is caused by a SQL statement that Access97
makes
involving NULL. The syntax that fails is "column_name" = NULL. The
following attachment was provided by -Jose'-. It contains a very small
enhancement to gram.y that will allow Access97 to work properly with sub
forms. Can this enhancement be added to release 6.5?
<<gram.patch>>
Thanks, Michael
the handling of negative numbers and commas. The attached path attempts
to fix these.
However the getValue method does not yet insert commas into the
generated string.
Also in getValue there is an incorrect assumption that the currency
symbol is '$', it should of course be '£'!. I have no idea on how to go
about fixing this one.
Alvin
I would like some feedback on what the hash function for the int8 hash
function
in the ./backend/access/hash/hashfunc.c should return.
Also, could someone (maybe Tomas Lockhart?) look-over the patch and make
sure
the system table entries are correct? I've tried to research them as
much as I
could, but some of them are still not clear to me.
Thanks,
-Ryan
and pg_operator. The lone error in pg_operator was reported as a bug
by Michael Reifenberger; the multiple errors in pg_proc would only have
been noticed if one invoked the functions by name rather than using
operator syntax. I guess few people do that.
palloc.h again. Move exporting of backend header files out of libpq's
Makefile (whatever was it doing there in the first place?) and into
backend/Makefile.
_copyResult didn't copy subPlan structure completely. _copyAgg is still
busted, apparently because of changes from EXCEPT/INTERSECT patch
(get_agg_tlist_references is no longer sufficient to find all aggregates).
No time to look at that tonight, however.
so remove them from MergeJoin node. Hack together a partial
solution for commuted mergejoin operators --- yesterday
a mergejoin int4 = int8 would crash if the planner decided to
commute it, today it works. The planner's representation of
mergejoins really needs a rewrite though.
Also, further testing of mergejoin ops in opr_sanity regress test.
+
+ Tue Feb 23 17:32:25 CET 1999
+
+ - Other than a struct a union itself cannot be specified as variable.
+
+ Fri Feb 26 07:18:25 CET 1999
+
+ - Synced preproc.y with gram.y.
+
+ Sat Feb 27 20:30:03 CET 1999
+
+ - Added automatic allocating for NULL pointers.
portability problem. Included patches should be applied to both
current and 6.4 tree. I have tested on LinuxPPC, FreeBSD and Solaris
2.6. Now the inet regression tests on these platforms are all happy.
---
Tatsuo Ishii
+
+ Son Feb 21 14:10:47 CET 1999
+
+ - Fixed variable detection in libecpg.
+
+ Mon Feb 22 19:47:45 CET 1999
+
+ - Added 'at <db_connection>' option to all commands it is apllicable
+ to. Due to changing the API of some libecpg functions this
+ requires me to increase the major version number.
+ - Synced pgc.l with scan.l.
+ - Added support for unions.
+ - Set library version to 3.0.0
+ - Set ecpg version to 3.0.0
file early enough to use in nodes/.
Try to be more complete for rules on generating parse.h,
but it still does not work any better than before. Should be able to
make correctly if parser/gram.y is updated even without a "make clean"
but so far not there yet.
shared memory space allocation. It's a wonder we have not seen bug
reports traceable to this area ... it's quite clear that the routine
dir_realloc() has never worked correctly, for example.
Ok. I made patches replacing all of "#if FALSE" or "#if 0" to "#ifdef
NOT_USED" for current. I have tested these patches in that the
postgres binaries are identical.
+
+ Fri Feb 19 21:40:14 CET 1999
+
+ - Fixed bug in libecpg that caused it to start transactions only for
+ the first connection.
+ - Set library version to 2.7.1
of MAXBACKENDS is now 1024, since all it's costing is about 32 bytes of memory
per array slot. configure's --with-maxbackends switch now controls DEF_MAXBACKENDS
which is simply the default value of the postmaster's -N switch. Thus,
the out-of-the-box configuration will still limit you to 64 backends,
but you can go up to 1024 backends simply by restarting the postmaster with
a different -N switch --- no rebuild required.
(--with-maxbackends). Add a postmaster switch (-N backends) that allows
the limit to be reduced at postmaster start time. (You can't increase it,
sorry to say, because there are still some fixed-size arrays.)
Grab the number of semaphores indicated by min(MAXBACKENDS, -N) at
postmaster startup, so that this particular form of bogus configuration
is exposed immediately rather than under heavy load.
Fix problem with date_part() for timespan (had an offset of one)
when given decade, century, and millenium as arguments.
Reported by Ricardo J.C.Coelho.
rule system semantics by having Var nodes referenced across multiple
parsetrees when rules split them.
Added more tests to the rules regression test.
The code in question resulted from v6.3 based development and was
a little careless applied to the v6.5 source tree.
Jan
qualification expression trees in the execution state. Prevents from
memory exhaustion on INSERT, UPDATE or COPY to tables that have CHECK
constraints. Speedup against the variant using freeObject() is more than
factor 2.
Jan
comparisons correctly. The psql monitor converts all table and field
names to lower case. If the PQfnumber function is called with a mixed
case name, it will always return -1.
Bahman Rafatjoo
for int8 support. configure now checks only snprintf() for int8 support,
not sprintf and sscanf as it used to. The reason for doing this is that
if we are supplying our own snprintf code (which does handle long long int),
we now only need working long long support in the compiler not in the
platform's C library. I have verified that int8 now passes regression test
on HPUX 9, and I think it should work on SunOS 4.1.* and other older
platforms if gcc is used.
I search in the planner for the '\xFF' appending.
Finally I found in MakeIndexable() in gram.y
Attach a patch which removes the "<=" test in USE_LOCALE,
might make some queries a bit slower for us "locale-heads",
BUT correct result is more important.
regards,
--
-----------------
Göran Thyni
o allow to use Big5 (a Chinese encoding used in Taiwan) as a client
encoding. In this case the server side encoding should be EUC_TW
o add EUC_TW and Big5 test cases to the regression and the mb test
(contributed by Jonah Kuo)
o fix mistake in include/mb/pg_wchar.h. An encoding id for EUC_TW was
not correct (was 3 and now is 4)
o update documents (doc/README.mb and README.mb.jp)
o update psql helpfile (bin/psql/psqlHelp.h)
--
Tatsuo Ishii
t-ishii@sra.co.jp
+
+ Wed Jan 27 12:42:22 CET 1999
+
+ - Fixed bug that caused ecpg to lose 'goto' information.
+ - Set ecpg version to 2.4.7
+
+ Fri Jan 29 18:03:52 CET 1999
+
+ - Fixed bug that caused 'enum' to be rejected in pure C code.
+ - Fixed bug that caused function names to be translated to lower case.
+ - Set ecpg version to 2.4.8
+
The following patch does two things.
- Clarifies what the effect of allowing users to add new users (Thet
become super-users.)
- Makes the default database for the new user if they are not allowed
to and the user agrees to create it.
Included patches fix a portability problem of unsetenv() used in
6.4.2 multi-byte support. unsetenv() is only avaliable on FreeBSD and
Linux so I decided to replace with putenv().
This implements some of the JDBC2 methods, fixes a bug introduced into the
JDBC1 portion of the driver, and introduces a new example, showing how to
use the CORBA ORB thats in Java2 with JDBC.
The Tar file contains the new files, the diff the changes to the others.
CHANGELOG is separate as I forgot to make a .orig ;-)
so that fetching an attribute value needs only one SearchSysCacheTuple call
instead of two redundant searches. This speeds up a large SELECT by about
ten percent, and probably will help GROUP BY and SELECT DISTINCT too.
was causing it not to detect out-of-range float values, as evidenced by
failure of float8 regression test. I corrected that logic and also
modified expected float8 results to account for new error message
generated for out-of-range inputs.
Pawel Pierscionek [pawel@astercity.net] reported about the
following case 1([SQL] drop table in pgsql).
Michael Contzen [mcontzen@dohle.com] reported about the
following case 2(PL/PGSQL bug using aggregates).
You can find it from pgsql-hackers archive.
1. PL/pgSQL can't execute UTILITY commands.
SPI_prepare() doesn't copy(save) the utilityStmt member of
Query type nodes,because copyObject() is not implemented
for nodes of (Create/Destroy etc)Stmt type.
2. Aggregates in PL/pgSQL cause wrong results.
...
It's a list including Aggreg type nodes which exist in
TargetList(i.e Aggreg type nodes are common to aggs
member list and TargetList).
AFAIC the common pointer is not copied to the same
pointer by copyObject() function.
In my patch I reconstruct aggs member node from
new(copied) Agg type node.
Is it proper to use set_agg_tlist_references() function to
reconstruct aggs member node for Agg type nodes ?
Thanks.
Hiroshi Inoue
Inoue@tpf.co.jp
elements prior to CREATEing new ones. It is under control of the -c
command line option (with the default being status quo).
The DROP TRIGGER portion still needs implementation. Anyone able to
help clarify what exactly the CREATE TRIGGER portion does so I can fix
this?
Again, I have tried this with tables/indexes/sequences, but do not
have other schema elements in my database. As a result, I am not 100%
convinced that I got the syntax correct in all cases (but think I did,
nonetheless). If anyone can check the other cases, I'd appreciate it.
Cheers,
Brook
[I added manual page and sgml additions for the new -c option.]
a field was labelled as a primary key, the system automatically
created a unique index on the field. This patch extends it so
that the index has the indisprimary field set. You can pull a list
of primary keys with the followiing select.
SELECT pg_class.relname, pg_attribute.attname
FROM pg_class, pg_attribute, pg_index
WHERE pg_class.oid = pg_attribute.attrelid AND
pg_class.oid = pg_index.indrelid AND
pg_index.indkey[0] = pg_attribute.attnum AND
pg_index.indisunique = 't';
There is nothing in this patch that modifies the template database to
set the indisprimary attribute for system tables. Should they be
changed or should we only be concerned with user tables?
D'Arcy
remaining shift/reduce conflict. But the very same conflict is in gram.y, so
I don't dig into it very much now.
Anyway, I just saw that there were minor changes made to ecpg by others. Now
I like that but I would prefer if I was told about that. Otherwise my
version numbering and Changelog maintaining might break. Or simply change
these too. :-)
Also I had to add #include <errno.h> to backend/libpq/pqcomprim.c to be
able to compile postgresql.
Patch is attached. Since my resubscription process is still not finished
yet, I still send them here.
Michael
INTERSECT and EXCEPT is available for postgresql-v6.4!
The patch against v6.4 is included at the end of the current text
(in uuencoded form!)
I also included the text of my Master's Thesis. (a postscript
version). I hope that you find something of it useful and would be
happy if parts of it find their way into the PostgreSQL documentation
project (If so, tell me, then I send the sources of the document!)
The contents of the document are:
-) The first chapter might be of less interest as it gives only an
overview on SQL.
-) The second chapter gives a description on much of PostgreSQL's
features (like user defined types etc. and how to use these features)
-) The third chapter starts with an overview of PostgreSQL's internal
structure with focus on the stages a query has to pass (i.e. parser,
planner/optimizer, executor). Then a detailed description of the
implementation of the Having clause and the Intersect/Except logic is
given.
Originally I worked on v6.3.2 but never found time enough to prepare
and post a patch. Now I applied the changes to v6.4 to get Intersect
and Except working with the new version. Chapter 3 of my documentation
deals with the changes against v6.3.2, so keep that in mind when
comparing the parts of the code printed there with the patched sources
of v6.4.
Here are some remarks on the patch. There are some things that have
still to be done but at the moment I don't have time to do them
myself. (I'm doing my military service at the moment) Sorry for that
:-(
-) I used a rewrite technique for the implementation of the Except/Intersect
logic which rewrites the query to a semantically equivalent query before
it is handed to the rewrite system (for views, rules etc.), planner,
executor etc.
-) In v6.3.2 the types of the attributes of two select statements
connected by the UNION keyword had to match 100%. In v6.4 the types
only need to be familiar (i.e. int and float can be mixed). Since this
feature did not exist when I worked on Intersect/Except it
does not work correctly for Except/Intersect queries WHEN USED IN
COMBINATION WITH UNIONS! (i.e. sometimes the wrong type is used for the
resulting table. This is because until now the types of the attributes of
the first select statement have been used for the resulting table.
When Intersects and/or Excepts are used in combination with Unions it
might happen, that the first select statement of the original query
appears at another position in the query which will be executed. The reason
for this is the technique used for the implementation of
Except/Intersect which does a query rewrite!)
NOTE: It is NOT broken for pure UNION queries and pure INTERSECT/EXCEPT
queries!!!
-) I had to add the field intersect_clause to some data structures
but did not find time to implement printfuncs for the new field.
This does NOT break the debug modes but when an Except/Intersect
is used the query debug output will be the already rewritten query.
-) Massive changes to the grammar rules for SELECT and INSERT statements
have been necessary (see comments in gram.y and documentation for
deatails) in order to be able to use mixed queries like
(SELECT ... UNION (SELECT ... EXCEPT SELECT)) INTERSECT SELECT...;
-) When using UNION/EXCEPT/INTERSECT you will get:
NOTICE: equal: "Don't know if nodes of type xxx are equal".
I did not have time to add comparsion support for all the needed nodes,
but the default behaviour of the function equal met my requirements.
I did not dare to supress this message!
That's the reason why the regression test for union will fail: These
messages are also included in the union.out file!
-) Somebody of you changed the union_planner() function for v6.4
(I copied the targetlist to new_tlist and that was removed and
replaced by a cleanup of the original targetlist). These chnages
violated some having queries executed against views so I changed
it back again. I did not have time to examine the differences between the
two versions but now it works :-)
If you want to find out, try the file queries/view_having.sql on
both versions and compare the results . Two queries won't produce a
correct result with your version.
regards
Stefan
file containing the latest version of the JDBC driver, allowing it to be
compiled and used under JDK 1.2 and later.
NB: None (well almost none) of the new methods actually do anything. This
release only handles getting it to compile and run. Now this is done, I'll
start working on implementing the new stuff.
Now this tar file replaces everything under src/interfaces/jdbc. I had to
do it this way, rather than diffs, because most of the classes under the
postgresql subdirectory have moved to a new directory under that one, to
enable the support of the two JDBC standards.
Here's a list of files in the tar file. Any file not listed here (in the
postgresql directory) will have to be deleted, otherwise it could cause
the driver to fail:
Peter Mount
seem to be portable (HPUX doesn't like it, anyway). Also, clean up
StreamConnection(), which was mis-coded to assume that the address
family field is already set when it's called.
Here's another patch for the libpq backend areas. This patch removes all
usage of "FILE *" on the communications channel. It also cleans up the
comments and headers in the pqcomm.c file - a lot of things were either
missing or incorrect. Finally, it removes a couple of unused functions
(leftovers from the time of shared code between the libpq backend and
frontend).
Here is a first patch to cleanup the backend side of libpq.
This patch removes all external dependencies on the "Pfin" and "Pfout" that
are declared in pqcomm.h. These variables are also changed to "static" to
make sure.
Almost all the change is in the handler of the "copy" command - most other
areas of the backend already used the correct functions.
This change will make the way for cleanup of the internal stuff there - now
that all the functions accessing the file descriptors are confined to a
single directory.
when deciding whether a field is a year field. Assume *anything* longer
than 2 digits (if it isn't a special-case doy) is a valid year.
This should fix the "Y1K" and "Y10K" problems
pointed out by Massimo recently.
Check usage of BC to require a positive-valued year; before just used it
to flip the sign of the year without checking. This led to problems
near year zero.
Allow a 5 digit "concatenated date" of 2 digit year plus day of year.
Do 2->4 digit year correction for 6 and 5 digit "concatenated dates".
Somehow forgot this originally. Guess not many folks use it...
I think NAN is already guaranteed to be there from Jan's work on NUMERIC,
but perhaps HUGE_VAL needs some #ifndef's in the same place.
Should also include "-Infinity" as -HUGE_VAL sometime; not there yet.
to give HAVE_TM_ZONE priority. This fixes glibc2 machines and any other
machine which passes both tests in configure.
Repair HAVE_TM_ZONE code which stuffs tm structure with date type values.
Same problems as were originally there before v6.1, but never noticed.
Thanks to Oleg for nagging :)
Nakajima. Since he is not subscribing the mailing list, I'm posting
his patches by his request. According to him, he has successfully
compiled and passed the regression test on Mac SE/30 running
NetBSD/m68k. Also, another person has reported that with the patches
PostgreSQL is working on NetBSD/sun3 too.
--
Tatsuo Ishii
+
+Wed Dec 9 11:24:54 MEZ 1998
+
+ - Synced preproc.y with gram.y and the keywords.c files to add CASE
+ statement.
+
+Tue Dec 22 14:16:11 CET 1998
+
+ - Synced preproc.y with gram.y for locking statements.
+ - Set version to 2.4.5
- the first patch is just to preven listing the perl warning in the
make output unless it is actually emitted by the make. this may
prevent new users from being confused by the warning in their output
- the second patch (to 2 files) just enables building/installing
pgaccess if TCL and TK are available. a Makefile is created to do
this, but you may wish to change the heading information in it since
I just copied another Makefile to use as a template.
I hope these make it into 6.4.1.
Cheers,
Brook
6.4.1. Here is the list:
- The type int8 now works. In fact, the bug(s) were in
src/backend/port/snprintf.c, so int8 is probably broken in every platform
that hasn't a native snprintf/vsnprintf. The type itself worked as
expected, only the output was wrong. Anyway, this patch should be checked
in other platforms.
- The regression tests for int2 and int4, which were broken due to
differences in the error messages, are fixed.
- The regression test for float8, which was broken in the reference
platform, is also fixed. I don't know if the new file (float8-OSF1.out)
will work on other platforms, but it might be worth to try it.
- Two new template files are provided (alpha_cc, which includes
optimization, and alpha_gcc), and src/templates/.similar is updated
accordingly. src/templates/alpha should be removed from the distribution.
*IMPORTANT NOTE*: I don't know if you can use gcc to compile postgres;
I've written the alpha_gcc file because alpha_cc has some flags that are
specific to DEC C.
- There is a (very basic) Digital Unix specific FAQ in
doc/FAQ_DigitalUnix.
--
-------------------------------------------------------------------
Pedro José Lobo Perea Tel: +34 91 336 78 19
missed before the release. It's simply a symbol that is undefined. This
patch defines this symbol in "win32.h", so it should have no effect on any
other platforms. It should go into 6.4.1 if possible, since compilation is
completely broken without it.
I am also attaching a patch for the "win32.mak" file - it leaves a file
behind when doing "make clean" after the library is built on Visual C++ 6.0.
This is not at all as urgent, but I don't see it breaking here, so I think
it might as well go in there too?
//Magnus
where you state a format and arguments. the old behavior required
each appendStringInfo to have to have a sprintf() before it if any
formatting was required.
Also shortened several instances where there were multiple appendStringInfo()
calls in a row, doing nothing more then adding one more word to the String,
instead of doing them all in one call.
support. Included patches will solve it and should be applied to
both trees. Also, it fix the problem with \c command of psql when
switching different encoding databases.
Regression tests passed.
--
Tatsuo Ishii
t-ishii@sra.co.jp
Fix one usage of substr() which mapped to the "Oracle compatibility" funcs
rather than the more recent (and closer to SQL92) function in varlena.c.
Add more DESC() entries for conversion functions.
unless necessary.
Label internal bpchar types as "character" and varchar types as
"character varying" to be less Postgres-specific. These types map to
the SQL92 definitions anyway.
Redefine g_force_quotes to be the local variable force_quotes.
Pass this as an argument to fmtId().
These should help with handling the single-byte internal "char" type.
over HAVE_INT_TIMEZONE. This may help out linux/glibc2 and Dec Alpha.
Included #error precompiler macros to catch cases where neither is defined
but USE_POSIX_TIME is (shouldn't happen). Hopefully this isn't just
a gcc-ism.
instead of our own halfway-there code. Add AC_STRUCT_TIMEZONE call
to check whether tm_zone exists in struct tm. Revise reading of template
file so that templates can define any variables they feel like (and,
indeed, can execute arbitrary shell code) rather than being constrained
to a fixed set of variable names.
destructions in 6.4 source using purify.
(1) parser/gram.y:fmtId()
It writes n+3 bytes into n+1 byte-long memory area if mixed case or
non-ascii identifiers given.
(2) catalog/index.c:
ATTRIBUTE_TUPLE_SIZE bytes are allocated but
sizeof(FormData_pg_attribute) bytes are written. Note that
ATTRIBUTE_TUPLE_SIZE is smaller than
sizeof(FormData_pg_attribute). (for example, on solaris 2.6,
Tatsuo Ishii
But it may be self-satisfied.
Please check my patch at the end of this posting.
Case 1. executor evaluates functions twice
Hiroshi Inoue
Inoue@tpf.co.jp
Digital Uni x with both DEC cc and gcc) behaviour of modifying an
lvalue on the left side an d then using it on the right side of an
assignment. Since this code modifies the
dbname parameter, it was changing, for example, "dbname=template1"
into "dbname =emplate1".
David Smith Programmer P
in the ACL code, and spell "GRANT RULE" correctly.
Apply patch from Oliver Elphick to not dump inherited constraints.
Apply patch from Constantin Teodorescu to dump table definitions with a
readable layout.
SunOS has tas(), but not memmove or strerror, and its sprintf() doesn't
return int. Also, older versions of GNU Make don't like rules with
empty left-hand sides...
instead of relying on port's os.h to tell us. (Needed for HPUX
where system major version is not enough info.)
configure unsets USE_TK if X libraries not found.
doc/Makefile uses gzcat or zcat as found by autoconf.
than silently returning zero on some machines. Correct float8 regress test
to agree. Also fix pow() overflow/underflow check to work correctly on
HPUX.
fail to consume the rest of the input string, and worse it would write
one more byte than it should into the buffer, probably resulting in coredump.
Fortunately there's a correct implementation next door in pqcomprim.c.
selected when they match a prefix of the value. The previous method,
which stripped all version data from and then tried to match that
against .similar entries, was entirely useless when .similar contained
several entries for different version numbers of a single OS name.
a backend core dump, because it was concatenating a potentially long
string onto another string that didn't necessarily have enough room.
Shame, shame.
This is the default, but the new flag will allow overriding an alias,
for example. So psql -n -N will put in the double quotes,
and psql -n can be an alias for psql.
Also, add a few braces around a nested single-line conditional construct
to suppress compiler warnings about "an ambiguous else".
From: SHIOZAKI Takehiko <takehi-s@ascii.co.jp>
I tried snapshot(Oct30) and made some patches.
# I think that it is confused to manage both Makefile.shlib and
# makefiles/Makefile.*, don't you?
* configure
Now FreeBSD 2.X is not supported..., so I added its entry.
If ELF_SYSTEM is set, gmake treat it defined even though
it is "false". So nothing should be set to use "ifdef".
BSD_SHLIB etc. may have same problems.
* Makefile.shlib
As you said, FreeBSD entry is much like BSD's.
I only added ELF_SYSTEM code.
* makefiles/Makefile.freebsd
Ifdef/else/endif can not be indented with TABs.
ie, not when user specifies --with-CC. This corrects a scripting error
that I'm surprised hasn't been reported more often. Moving the macro call
to the earlier point in the script is correct anyway: if -traditional is needed,
it should get added to CFLAGS before we start using the compiler for
other tests.
mistakes in creating pg_operator table. NOTE: right now, this will
fail because of conflicting definitions for point @ path operator.
I trust we're gonna fix that.
oprlsortop and oprrsortop links. There's still a bug involving
conflicting definitions for point @ path, but I'm not taking
responsibility for deciding which one is right...
(Someone forgot whether their subroutine signaled errors by a NULL pointer
return value, or a negative integer... I'm surprised gcc -Wall doesn't
catch this...)
Fixes a bug in the rule system that caused a crashing
backend when a join-view with calculated column is used
in subselect.
Modifies EXPLAIN to explain rewritten queries instead of
the plain SeqScan on a view. Rules can produce very deep
MORE
Jan.
src/Makefile.shlib. Updated all the makefiles that try to build shlibs
to include that file instead of having duplicate (and mostly incomplete)
copies of shared-library options. It works on HPUX, a lot better than it
did before in fact, but there's a chance I broke some other platforms.
At least now you only have to fix one place not six...
Get the permissions right, don't overwrite real files with symlinks, etc.
plpgsql and odbc still aren't fully up to speed, but at least they don't crash and burn...
problem:
'tclsh' still had to be found even if --with-libs (or
--with-libraries) was
specified to configure.
--with-libs is really an overloaded option. It really should only be used
to specify additions directories to search in order to file needed
libraries. It was also being used to locate the *Config.sh files.
Billy G. Allie
libtcl has been installed as a non-shared library. pltcl cannot be
built in that situation; we want to do nothing and let the overall Postgres
build complete, rather than failing.
DataDir is set after read_pg_options if postgres is called
interactively. If postgres is forked by postgres DataDir is read from
the PGDATA enviromnent variable set by the postmaster and this explains
while the bug disappears. I have written this patch but I don't like
it. Any better idea?
Massimo Dal Zotto
compile out of the tar file on Solaris with the SUN 5.0 compilers.
These compilers will be needed if you wan to compile the libpg++
interface without using the gcc/g++. The SC4.2 compilers do not
understand the string class.
The first patch changes the ecpg intermediate shared library
name from *.sho to *.sho.o so that the SUN compiler will
allow it to be used in conjunction with the -o option.
Matthew C. Aycock
Here are patches needed to complie under AIX 4.2.
I changed configure.in, pqcomm.c, config.h.in, and fe-connect.c.
Also I had to install flex because lex did not want to translate pgc.l.
do not configure in the perl5 interface.
the perl5 interface needs to be installed under /usr/local/lib/perl5/*, which
is generally owned by root. This allows a non-root build/install with the
only root requirement being the make/install of hte perl5 stuff...
When importing an image into the database, the example now fires off a
new
Thread, which imports the image in the background. This also means that
the application doesn't freeze on the user, and they can still browse
the
images in the database, while the upload is running.
This now makes the ImageViewer a true example on how to use Threads (the
threadtest class is just that - a test).
Peter
following patches fix the problems (i.e., all regression tests pass)
in what I hope to be a platform-independent fashion. The accomplish
the following:
Brook Milligan
is wrong and dangerous unless you are using contrib/string. We really
need a thorough look at the issue of making the backend and the FE/BE
protocols completely 8-bit-clean for string data, but that's a task
for some future release.
newly-updated SGML reference pages, so I just inserted a comment that they
are obsolete. If you want to transcribe the newer info into these pages,
be my guest.
important step towards making the driver compliant, and means that for
some Java applications and servlets, only a single database connection
is
needed, so in a sence this is a nice little show stopper for 6.4 (and
should still be backward compatible to 6.3.2).
Peter
Here are two new patches for the Win32 support.
1) The patch based on the one from Hiroshi Inoue [Inoue@tpf.co.jp], to
load
Winsock.dll from libpq.dll.
2) A patch for psql.c to remove the call to WSAStartup(), since it is
not
required when it's done in libpq.dll.
I'm still looking for the possibility of having a crypt() function in
libpq.dll too, the same way getopt was included. Any chance of getting
this
before 6.4, or should we wait for the next one?
//Magnus
Use @top-srcdir@ to find the right Makefile.global and use ODBCSRCDIR
to point to this local directory.
Move non-platform-specific stuff to outside the if clauses.
Still need to move all platform-specific stuff to the templates.
Before, "make install" did not run the lextest.
Fix up the ODBC make from this main configure.
Include configure test for "ln -s" in Makefile.global.in.
Was always in configure, just not carried through to here for use.
to get rid of unused variables.
Get clean compile on Linux (Thomas and Gerald).
Implement autoconf/configure for standalone builds and use the existing
autoconf/configure system when in the Postgres source tree.
Code tests and functions with ApplixWare-4.4.1beta on a Linux box.
Changes should be backward compatible with WIN32 but still needs testing.
Is it too late to add a feature to pg_dump for 6.4??
I just spent most of the day learning pg_dump and modifing it so it
would
dump views also.
This is the first time I have ever contributed any code changes, so I'm
not sure of how to submit it.
The diff's and a readme as a tgz file are attached.
Thanks
Terry Mackintosh <terry@terrym.com> http://www.terrym.com
regression test on a FreeBSD box with both non-MULTIBYTE and
MULTIBYTE-enabled, and confirmed that the results are same.
However I do not tested on PCs(I don't have access to win). Please let
me know if the patches break anything on PCs.
Also please note that the patch for varchar.c is a fix for a nasty bug
of char(n) types that I introduced and I believe at least this should
be applied.
Tatsuo Ishii
and make backend/libpq/pqcomm.c only try to lock the socket file when
the call exists. Also, change open-RDONLY to open-WRONLY; at least
on my platform, you can't get a write lock on a file you didn't open
for writing.
Check strdup calls for out of memory.
Set library version to 2.6.2
Synced preproc.y and keywords.c with gram.y and keywords.c yet again.
Set version to 2.4.3
columns of views at all (not only oid, cmin etc. too).
pgsql=> select cmin from pg_rules;
ERROR: system column cmin not available - pg_rules is a view
pgsql=> select * from pg_rules where pg_rules.oid = pg_class.oid;
ERROR: system column oid not available - pg_rules is a view
pgsql=>
Jan
Formerly did so only for those which clearly required it, but that
would still miss things like reserved key words which also require it.
Implement the "-n" switch to revert the double quote behavior
to put DQs only where there is more than lower-case, digits,
and underscores.
for against a just updated CVS tree. It contains
Partial new rewrite system that handles subselects, view
aggregate columns, insert into select from view, updates
with set col = view-value and select rules restriction to
view definition.
Updates for rule/view backparsing utility functions to
handle subselects correct.
New system views pg_tables and pg_indexes (where you can
see the complete index definition in the latter one).
Enabling array references on query parameters.
Bugfix for functional index.
Little changes to system views pg_rules and pg_views.
The rule system isn't a release-stopper any longer.
But another stopper is that I don't know if the latest
changes to PL/pgSQL (not already in CVS) made it compile on
AIX. Still wait for some response from Dave.
Jan
parameters. With it applied a function like
CREATE FUNCTION getname(oid8, int4) RETURNS name AS
'SELECT typname FROM pg_type WHERE oid = $1[$2]'
LANGUAGE 'sql';
is possible. Mainly I need this to enable array references in
expressions for PL/pgSQL. Complete regression test ran O.K.
Jan
spin-locks. Notice that it's now inline assembler in s_lock.h,
rather than seperate code in s_lock.c. It also shrank a little
bit... Just rip out the S_LOCK() define and insert the tas() inline
function. Please let me know if there are any problems with it.
Jon Buller
and what wasn't. Also try to improve the comments so that doesn't happen
again. Changed SIGPIPE handling to SIG_IGN so that if frontend quits,
we will finish out the current command and return to main loop before
quitting. This seems much safer than a forced abort mid-command.
Add "timestamp" to list of tokens in keywords.c.
Before, TIMESTAMP WITH TIME ZONE did not actually parser.
Reorder token lists to be more alphabetical.
Remove ARCHIVE keyword which was deprecated in v6.3.
here is the patch that includes PL/pgSQL into the build
(currently with make errors ignored) and adds a regression
test for it. A clean build and regression ran fine here.
Can you please apply it?
The tar should be extracted in /usr/local/src/pgsql and
creates the following files:
src/pl/Makefile
called by toplevel GNUmakefile and for now only calls
src/pl/plpgsql/Makefile
src/pl/plpgsql/Makefile
calls src/pl/plpgsql/src/Makefile (here the call to
make ignores build errors - this must be changed
later for the final release).
src/test/regress/input/install_plpgsql.source
SQL script installing PL/pgSQL language in regression
database. Will be modified by .../input/Makefile to
point to correct PGLIB directory where plpgsql.so
gets installed.
src/test/regress/output/install_plpgsql.source
expected output for installation script.
src/test/regress/sql/plpgsql.sql
the main regression test. It tests functions and
triggers written in PL/pgSQL including views that use
supportfunctions in this language.
src/test/regress/expected/plpgsql.out
the expected output for the above regression test.
make_plpgsql.diff
patch that adds some lines to
src/GNUmakefile.in
src/test/regress/expected/Makefile
src/test/regress/input/Makefile
src/test/regress/output/Makefile
src/test/regress/sql/Makefile
src/test/regress/sql/tests
test passes. Interestingly, the fix involves no changes or special
cases in the union test and actually removes a special case for the
numerology test. Thus, following the strategy outlined below is a
definite improvement over the previous situation.
Cheers,
Brook
+ Mon Aug 31 09:40:04 CEST 1998
+
+ - Minor patch to Makefile
+ - Put pgc.l in sync with scan.l
+
+ Tue Sep 1 11:31:05 CEST 1998
+
+ - Fixed another bug in preproc.y
+
+ Thu Sep 3 12:21:16 CEST 1998
+
+ - Sync preproc.y with gram.y
+
+ Mon Sep 14 09:21:02 CEST 1998
+
+ - Sync preproc.y with gram.y yet again
+
+ Thu Sep 17 08:55:33 CEST 1998
+
+ - Synced preproc.y and gram.y one more time
+
+ Thu Sep 17 19:23:24 CEST 1998
+
+ - Added missing ´;´ in preproc.y
+ - Set version to 2.4.2
* It works under both Tcl 7.6 and Tcl 8.0 now. (The code claims to
work under Tcl 7.5 as well, but I have no way to test that ---
if anyone still cares, please check it with 7.5.)
* pg_listen suppresses extra LISTEN commands and correctly sends an
UNLISTEN when the last listen request for a relation is cancelled.
(Note this means it will not work with pre-6.4 backends, but that
was true already because it depends on the current libpq, which
only speaks protocol 2.0.)
* Added -error option to pg_result so that there's some way to find
out what you did wrong ;-)
* Miscellaneous cleanups of code comments and overenthusiastic #includes.
BTW, I bumped the package version number from 1.2 to 1.3. Is this
premature? Does someone run around and do that routinely before
each pgsql release?
regards, tom lane
prompt_for_password code that psql does. We fixed psql a month or
two back to permit usernames and passwords longer than 8 characters.
I propagated the same fix into pg_dump.
Tom Lane
Change DEFAULT NULL to send back a NULL pointer
rather than a string "NULL". This seems to work, where sending
the string led to type conversion problems (and probably the wrong
thing anyway).
rather than func_select_candidate().
Fix oper_select_candidate() to work with a single operator argument.
Repair left operator checking for null return from candidate list.
as, if not installing as root (which nobody *should* be doing, of course),
the perl install fails, which means that both bin and man directories
are not installed :(
This way, only thing that doesn't get installed is perl interface...
compiler to
attempt to compile libpq++. The patches address the following problems:
1. In my first pass at changing the libpq++ makefile, I forgot to
include the
PORTNAME in the Makefile.in file.
2. The UnixWare 7 C++ compiler did not like the '-K alloca' option in
CXXFLAGS.
Billy G. Allie
implementations of strtol() treat empty strings ("") as invalid arguments
while others convert this (erroneously, IHMO) to zero (0). Assuming that the
expected behaviour of pg_atoi() is to return 0 if it is passed an empty
string, I am supplying the following patch to explictly check for an empty
string in pg_atoi() and return 0 if the string is empty. The patch will also
trap a NULL character pointer being passed to pg_atoi() and will use elog() to
print out an error message if the input char pointer is NULL.
Billy G. Allie
1. The UnixWare tas macro was reformatted (by indent or it like?) which caused
it to break. The asm macro construct is very particular about the %mem
construct -- it has to start in column 1.
2. When compiling libpq++, g++ was used even if configure found the C++ com-
piler to be CC.
3. When compiling libpq++, '-Wno-error' was added to CXXFLAGS, even if the
compiler wasn't g++.
Billy G. Allie
> Open portability issues:
>
> /usr/local should be searched for lib and include for all ports if
present
> (currently not working, I have libreadline there)
>
> the stream functions on AIX need a size_t for addrlen's in
fe-connect.c and pqcomm.c.
>
> lock.c still has an incompatible TPRINTF(flags, args...) definition
Massimo
Here's a patch for initdb that does two things.
1) Encloses the created rulenames in quotes to preserve case
in the creation step. (stores _RETpg... instead of _retpg...)
I believe _RET is standard for views.
2) Renames pg_view to pg_views and pg_rule to pg_rules.
I believe Jan and myself agreed this would be a "good idea"
Keith Parks
I put some extra checks to make sure a query was a good candidate for
rewrite into a UNION. Besides the existing checks:
1. Make sure the AND/OR tree was rectangular. ( i.e. 3 X 4 or 10 X
3)
2. Only one table.
3. Must have an AND dimension.
4. At least 9 OP expressions total
Also cleaned up and commented.
before.
> Looks like a GNU-ism. I nice one, but still a GNU-ism.
Sorry, I didn't know it is a GNU extension. I have written this patch
which should fix the problem. Let me know if you still have problems.
Massimo Dal Zotto
compiled with -O0. Included are patches that should fix the problem
(of course I have confirmed -O2 works with this patch).
BTW, here is a platforms/regression test failure(serious one--backend
death) matrix.
Tatsuo Ishii
structs from libpq-fe.h, as we previously discussed.
There turned out to be sloppy coding practices in more places than
I had realized :-(, but all in all I think it was a well-worth-while
exercise.
I ended up adding several routines to libpq's API in order to respond
to application requirements that were exposed by this work. I owe the
docs crew updates for libpq.sgml to describe these changes. I'm way too
tired to work on the docs tonight, however.
This is the last major change I intend to submit for 6.4. I do want
to see if I can make libpgtcl work with Tcl 8.0 before we go final,
but hopefully that will be a minor bug fix.
>
> Please apply this HAVING regression patch.
> > My bad. It is caused by a known bug having to do with GROUP BY.
It ain't$
> > nothing to do with HAVING. For some reason the bug went away for a
while, $
> > script. It must have, because that is how I created the expected
file. :(
> >
> > A patch to the regression will be forthcoming.
>
We're carrying around a copy of install-sh in case the local system
has no install script. It's wasted baggage, because configure doesn't
know it's there :-(. (Apparently everyone who's used postgres lately
already had an install script somewhere in their path. I happened to
try to run configure with a minimal PATH tonight, and it promptly
gave up for lack of an install program.) Here's the patch.
After some playing with gdb I found that in printtup() there is a non null
attribute with typeinfo->attrs[i]->atttypid = 0 (invalid oid). Unfortunately
attibutes with invalid type are neither printed nor marked as null, and this
explains why psql doesn't get all the expected data.
So I made this patch to printtup():
ODBC driver have found a bug in 6.3.2 pg_dump and have made patches.
I confirmed that the same bug still exists in the current source
tree. So I made up patches based on Kataoka's. Here are some
explanations.
o fmtId() returns pointer to a static memory in it. In the meantime
there is a line where is fmtId() called twice without saving the
first value returned by fmtId(). So second call to fmtId() will
break the first one.
o findTableByName() looks up a table by its name. if a table name
contanins upper letters or non ascii chars, fmtId() will returns
a name quoted in double quotes, which will not what findTableByName()
wants. The result is SEG fault. -- Tatsuo Ishii t-ishii@sra.co.jp
Here is a new patch for libpq, to make it work on Win32 again (since
the latest modifications broke it a little).
Please also add the file "libpq.rc" to the interfaces/libpq directory.
This will allow version-stamping of the generated DLL file, so that
automatic install programs (and interested users) can determine
the version of the file. The file is currently set as "prerelease".
Before the release, somebody should change the line "FILEFLAGS
VS_FF_PRERELEASE" to "FILEFLAGS 0". That information should probably
go into toos\RELEASE_CHANGES.
The patch is against the cvs as of ~ 1998-08-26 14:30 CEST.
//Magnus
This one is against the current archive (so it contains the one I send the
other day). It should fix the AIX problems. Andreas, could you please try
it? Thanks.
+ Wed Aug 26 16:17:39 CEST 1998
+
+ - Sync preproc.y with gram.y
+
+ Thu Aug 27 15:32:23 CEST 1998
+
+ - Fix some minor glitches that the AIX compiler complains about
+ - Added patchlevel to library
+
+ Fri Aug 28 15:36:58 CEST 1998
+
+ - Removed one line of code that AIX complains about since it was not
+ needed anyway
+ - Set library version to 2.6.1
I don't know if this is really related to the initdb problem
discussion (haven't followed it enough). But seems so because
it fixes a damn problem during index tuple insertion on
CREATE TABLE into pg_attribute_relid_attnum_index.
Anyway - this bug was really hard to find. During startup the
relcache reads in some prepared information about index
strategies from a file and then reinitializes the function
pointers inside the scanKey data. But for sake it assumed
single attribute index tuples (hasn't that changed recently).
Thus not all the strategies scanKey entries where initialized
properly, resulting in invalid addresses for the btree
comparision functions.
With the patch at the end the regression tests passed
excellent except for the sanity_check that crashed at vacuum
and the misc test where the select unique1 from onek2 outputs
the two rows in different order.
Jan
Ok. Here is a patch to make psql work on Win32 (as a console mode
application, of course).
It requires getopt.c to be in src/utils - works fine with the FreeBSD
version of it.
Also, the file win32.mak should go into src/bin/psql.
> these patches define the UNLISTEN sql command. The code already
> existed but it was unknown to the parser. Now it can be used
> like the listen command.
> You must make clean and delete gram.c and parser.h before make.
> tprintf.patch
>
> tprintf.patch
>
> adds functions and macros which implement a conditional trace package
> with the ability to change flags and numeric options of running
> backends at runtime.
> Options/flags can be specified in the command line and/or read from
> the file pg_options in the data directory.
> socket-flock.patch
>
> use advisory locks to check if the unix socket can be deleted.
> A running postmaster keeps a lock on that file. A starting
> postmaster exits if the file exists and is locked, otherwise
> it deletes the sockets and proceeds.
> This avoid the need to remove manually the file after a postmaster
> or system crash.
> I don't know if flock is available on any system. If not we could
> define a HAVE_FLOCK set by configure.
> sinval.patch
>
> fixes a problem in SI cache which causes table overflow if some
> backend is idle for a long time while other backends keep adding
> entries.
> It uses the new signal handling implemented in tprintf.patch.
> I have also increacasesed the max number of backends from 32 to 64
> and the table size from 1000 to 5000.
> I don't know if anybody is working on SI, but until another
> solution is found this patch fixes the problem. I have received
> messages from other people reporting the same problem which I
> fixed many months ago.
> sequence.patch
>
> adds the missing setval command to sequences. Owner of sequences
> can now set the last value to any value between min and max
> without recreating the sequence. This is useful after loading
> data from external files.
> ps-status.patch
>
> macros for ps status, used by postgres.c and utility.c.
> Unfortunately ps status is system dependent and the current
> code doesn't work on linux. The use of macros confines system
> dependency to into one file (ps-status.h). Users of other
> operating systems should check this code and submit new macros.
lock.patch
I have rewritten lock.c cleaning up the code and adding better
assert checking I have also added some fields to the lock and
xid tags for better support of user locks. There is also a new
function which returns an array of pids owning a lock.
I'm using this code from over six months and it works fine.
assert.patch
adds a switch to turn on/off the assert checking if enabled at compile
time. You can now compile postgres with assert checking and disable it
at runtime in a production environment.
(Mark or Bruce?) It fixes a problem when cpp gives a warning when
precompiling /dev/null like: "/dev/null", line 1: 1506-229 (W)
File is empty. This leads to a hangup when doing the description
load during initdb, since stderr also ends up in the global1.description
and local1_template1.description
stderr has to be redirected to /dev/null:
Andreas Zeugswetter
statements:
- the table definition with a default clause referencing the sequence;
- a CREATE SEQUENCE statement;
- a UNIQUE constraint, which expands into a CREATE INDEX statement.
This is not a perfect solution, since the sequence will remain even if
the table is dropped. Also, there is no absolute protection on updating
the sequence column.
+
+ Fri Aug 14 12:44:21 CEST 1998
+
+ - Added EXEC SQL DEFINE statement
+ - Set version to 2.4.0
+
+ Tue Aug 18 09:24:15 CEST 1998
+
+ - Removed keyword IS from DEFINE statement
+ - Added latest changes from gram.y
+ - Removed duplicate symbols from preproc.y
+ - Initialize sqlca structure
+ - Added check for connection to ecpglib
+ - Set version to 2.4.1
+
+ Thu Aug 20 15:31:29 CEST 1998
+
+ - Cleaned up memory allocation in ecpglib.c
+ - Set library version to 2.6
+
+
+ Fri Aug 14 12:44:21 CEST 1998
+
+ - Added EXEC SQL DEFINE statement
+ - Set version to 2.4.0
+
+ Tue Aug 18 09:24:15 CEST 1998
+
+ - Removed keyword IS from DEFINE statement
+ - Added latest changes from gram.y
+ - Removed duplicate symbols from preproc.y
+ - Initialize sqlca structure
+ - Added check for connection to ecpglib
+ - Set version to 2.4.1
+
+ Thu Aug 20 15:31:29 CEST 1998
+
+ - Cleaned up memory allocation in ecpglib.c
+ - Set library version to 2.6
+
+
+ Fri Aug 14 12:44:21 CEST 1998
+
+ - Added EXEC SQL DEFINE statement
+ - Set version to 2.4.0
+
+ Tue Aug 18 09:24:15 CEST 1998
+
+ - Removed keyword IS from DEFINE statement
+ - Added latest changes from gram.y
+ - Removed duplicate symbols from preproc.y
+ - Initialize sqlca structure
+ - Added check for connection to ecpglib
+ - Set version to 2.4.1
+
+ Thu Aug 20 15:31:29 CEST 1998
+
+ - Cleaned up memory allocation in ecpglib.c
+ - Set library version to 2.6
+
Thanks. But patches for src/backend/catalog/Makefile seems missing
in the current source tree. Please apply attached patches.
It also includes some corrections to src/backend/util/mb/wchar.c.
-- Tatsuo Ishii t-ishii@sra.co.jp
[AC_MSG_RESULT(yes) AC_DEFINE(HAVE_LONG_INT_64)],
this line produces something like:
echo "$ac_t""yes" 1>&6 cat >> confdefs.h <<\EOF
and would append garbage "yes cat" to confdefs.h. Of course the
result confdefs.h is not syntactically correct therefore following
tests using confdefs.h would all fail. To avoid the problem, we
could switch the order of AC_MSG_RESULT and AC_DEFINE (see attached
patch). This happend on my LinuxPPC box.
Tatsuo Ishii t-ishii@sra.co.jp
patch is applied:
Rewrite rules on relation level work fine now.
Event qualifications on insert/update/delete rules work
fine now.
I added the new keyword OLD to reference the CURRENT
tuple. CURRENT will be removed in 6.5.
Update rules can reference NEW and OLD in the rule
qualification and the actions.
Insert/update/delete rules on views can be established to
let them behave like real tables.
For insert/update/delete rules multiple actions are
supported now. The actions can also be surrounded by
parantheses to make psql happy. Multiple actions are
required if update to a view requires updates to multiple
tables.
Regular users are permitted to create/drop rules on
tables they have RULE permissions for
(DefineQueryRewrite() is now able to get around the
access restrictions on pg_rewrite). This enables view
creation for regular users too. This required an extra
boolean parameter to pg_parse_and_plan() that tells to
set skipAcl on all rangetable entries of the resulting
queries. There is a new function
pg_exec_query_acl_override() that could be used by
backend utilities to use this facility.
All rule actions (not only views) inherit the permissions
of the event relations owner. Sample: User A creates
tables T1 and T2, creates rules that log
INSERT/UPDATE/DELETE on T1 in T2 (like in the regression
tests for rules I created) and grants ALL but RULE on T1
to user B. User B can now fully access T1 and the
logging happens in T2. But user B cannot access T2 at
all, only the rule actions can. And due to missing RULE
permissions on T1, user B cannot disable logging.
Rules on the attribute level are disabled (they don't
work properly and since regular users are now permitted
to create rules I decided to disable them).
Rules on select must have exactly one action that is a
select (so select rules must be a view definition).
UPDATE NEW/OLD rules are disabled (still broken, but
triggers can do it).
There are two new system views (pg_rule and pg_view) that
show the definition of the rules or views so the db admin
can see what the users do. They use two new functions
pg_get_ruledef() and pg_get_viewdef() that are builtins.
The functions pg_get_ruledef() and pg_get_viewdef() could
be used to implement rule and view support in pg_dump.
PostgreSQL is now the only database system I know, that
has rewrite rules on the query level. All others (where I
found a rule statement at all) use stored database
procedures or the like (triggers as we call them) for
active rules (as some call them).
Future of the rule system:
The now disabled parts of the rule system (attribute
level, multiple actions on select and update new stuff)
require a complete new rewrite handler from scratch. The
old one is too badly wired up.
After 6.4 I'll start to work on a new rewrite handler,
that fully supports the attribute level rules, multiple
actions on select and update new. This will be available
for 6.5 so we get full rewrite rule capabilities.
Jan
the following to regress/sql/tests.
If applying by hand note that the setup_... must run before
the run_... (that I splitted these two was due to the errors
that occured when creating rules and using them then in the
same session - I'll post another fix for this later).
BTW: the regression tests sanity_checks and alter_table fail
now due to the remove of some indices and the oidint4 and
oidname types. At least expectes should be set to the current
results.
Thanks.
Jan
if MULTIBYTE is not enabled. So be sure to run initdb.
o these patches are made against the latest source tree (after
Bruce's massive patch, I think) BTW, I noticed that after running
regression, the oid field of pg_type seems disappeared.
regression=> select oid from pg_type; ERROR: attribute
'oid' not found
this happens after the constraints test. This occures with/without
my patches. strange...
o pg_database_mb.h, pg_class_mb.h, pg_attribute_mb.h are no longer
used, and shoud be removed.
o GetDatabaseInfo() in utils/misc/database.c removed (actually in
#ifdef 0). seems nobody uses.
t-ishii@sra.co.jp
is a working 64-bit-int type available.
In playing around with it on my machine, I found that gcc provides
perfectly fine support for "long long" arithmetic ... but sprintf()
and sscanf(), which are system-supplied, don't work :-(. So the
autoconf test program does a cursory test on them too.
If we find that a lot of systems are like this, it might be worth
the trouble to implement binary<->ASCII conversion of int64 ourselves
rather than relying on sprintf/sscanf to handle the data type.
regards, tom lane
usernames and passwords work correctly in both "password" and
"crypt" authorization mode. NOTE: at least on my machine, it seems
that the crypt() routines ignore the part of the password beyond
8 characters, so there's no security gain from longer passwords in
crypt auth mode. But they don't fail.
The login-related part of psql has apparently not been touched
since roughly the fall of Rome ;-). It was going through huge
pushups to get around the lack of username/login parameters to
PQsetdb. I don't know when PQsetdbLogin was added to libpq, but
it's there now ... so I was able to rip out quite a lot of crufty
code while I was at it.
It's possible that there are still bogus length limits on username
or password in some of the other PostgreSQL user interfaces besides
psql/libpq. I will leave it to other folks to check that code.
regards, tom lane
with the new support for asynchronous NOTIFY in libpgtcl. With
the current sources, if the backend disconnects unexpectedly then
the tcl/tk application coredumps when control next reaches the idle
loop. Oops.
regards, tom lane
Summary of changes:
In pqcomm.h, use the SUN_LEN macro if it is defined to calculate
the size of the sockaddr_un structure.
In unixware.h, drop the use of the UNIXWARE macro. Everything can
be handled with the USE_UNIVEL_CC and DISABLE_COMPLEX_MACRO macros.
In s_lock.h, remove the reference to the UNIXWARE macro (see above).
In the unixware template, add the YFLAGS:-d line.
In various makefile templates, add (or cleanup) unixware and univel
port specific information.
-- Billy G. Allie
no longer returns buffer pointer, can be gotten from scan;
descriptor; bootstrap can create multi-key indexes;
pg_procname index now is multi-key index; oidint2, oidint4, oidname
are gone (must be removed from regression tests); use System Cache
rather than sequential scan in many places; heap_modifytuple no
longer takes buffer parameter; remove unused buffer parameter in
a few other functions; oid8 is not index-able; remove some use of
single-character variable names; cleanup Buffer variables usage
and scan descriptor looping; cleaned up allocation and freeing of
tuples; 18k lines of diff;
in constraint clauses.
IN and NOT IN only allow constaints, not subselects.
Jose' Soares' new reference docs pointed out the discrepency.
Updating the docs too...
Sigh. That tweak needs a tweak --- I didn't realize that ".DEFAULT"
processing ignores dependencies, at least in the version of gmake I
have here (not sure if it's a bug or not). Apply this patch aftermy previous one...
From: Tom Lane <tgl@sss.pgh.pa.us>
Attached is a patch for this weekend's work on libpq. I've dealt
with several issues:
<for details: see message, in pgsql-patches archive for above data>
Bracket things with #ifdef ENABLE_LINE_TYPE.
The line data type has always been used internally to support other types,
but I/O routines have never been defined for it.
indices for restriction clauses containing a constant.
Note that if an index does not match directly (usually because the types
on both side of the clause don't match), and if a binary-compatible index
is identified, then the operator function will be replaced by a new
one. Should not be a problem, but be sure that if types are listed as
being binary compatible (in parse_coerce.h) then the comparison functions
are also binary-compatible, giving equivalent results.
trouble, and the name of the shared library has been changed recently.
Had to rerun ldconfig on my machine to get it working again.
Give an error message with a helpful hint if so...
functions btrim() ltrim() and rtrim().
The error was that the character after the set was included
in the tests (ptr2 pointed to the character after the vardata
part of set if no match found, so comparing *ptr or *end
against *ptr2 MAY match -> strip).
Jan
--
#======================================================================#
# It's easier to get forgiveness for being wrong than for being
right. # # Let's break this rule - forgive me.
# #======================================== jwieck@debis.com (Jan
Wieck) #
couple weeks ago on the hackers and interfaces lists:
1. When the backend sends a NOTICE message and closes the connection
(typically, because it was told to by the postmaster after
another backend coredumped), libpq will now print the notice
and close the connection cleanly. Formerly, the frontend app
would usually terminate ungracefully due to a SIGPIPE. (I am
not sure if 6.3.2 behaved that way, but the current cvs sources
do...)
2. libpq's various printouts to stderr are now fed through a single
"notice processor" routine, which can be overridden by the
application to direct notices someplace else. This should ease
porting libpq to Windows.
I also noticed and fixed a problem in PQprint: when sending output
to a pager subprocess, it would disable SIGPIPE in case the pager
terminates early (this is good) --- but afterwards it reset SIGPIPE
to SIG_DFL, rather than restoring the application's prior setting
(bad).
regards, tom lane
I have attached a patch to allow GROUP BY and/or ORDER BY function or
expressions. Note worthy items:
1. The expression or function need not be in the target list.
Example:
SELECT name FROM foo GROUP BY lower(name);
2. Simplified the grammar to use expressions only.
3. Cleaned up earlier patch in this area to make use of existing
utility functions.
3. Reduced some of the members in the SortGroupBy parse node. The
original data members were redundant with the new expression node.
(MUST do a "make clean" now)
4. Added a new parse node "JoinUsing". The JOIN USING clause was
overloading this SortGroupBy structure. With the afore mentioned
reduction of members, the two clauses lost all their commonality.
5. A bug still exist where, if a function or expression is GROUPed BY,
and an aggregate function does not include a attribute from the
expression or function, the backend crashes. (or something like
that) The bug pre-dates this patch. Example:
SELECT lower(a) AS lowcase, count(b) FROM foo GROUP BY lowcase;
*** BOOM ***
--Also when not in target list
SELECT count(b) FROM foo GROUP BY lower(a);
*** BOOM AGAIN ***
There are three or four cases in transformSortClause() and I had fixed
only one case for UNION. A second case is now fixed, in the same way; I
assigned INT4OID to the column type for the "won't actually happen"
sort. Didn't want to skip the code entirely, since the backend needs to
_try_ a sort to get the NULLs right. I'm not certain under what
circumstances the other cases are invoked and these are not yet
fixed up, though perhaps they don't need to be...
least, Solaris 2.5.1. We use it in backend/utils/adt/int8.c.
Add a check to configure so that we see if it exists or not, and, if not,
compile in snprintf.c from backend/port, which was taken from, and falls under
the same Berkeley license as us, the FreeBSD libc/stdio ...
As Bruce mentioned, this is due to the conflict among changes we made.
Included patches should fix the problem(I changed all MB to
MULTIBYTE). Please let me know if you have further problem.
P.S. I did not include pathces to configure and gram.c to save the
file size(configure.in and gram.y modified).
Everyone using an [NOT] EXISTS subquery will have noticed that
already.
The bug is in "subselect.c" in the function "SS_process_sublinks()".
Here the whole function as it *SHOULD BE*:
Stephan
From: t-ishii@sra.co.jp
Attached are patches to enhance the multi-byte support. (patches are
against 7/18 snapshot)
* determine encoding at initdb/createdb rather than compile time
Now initdb/createdb has an option to specify the encoding. Also, I
modified the syntax of CREATE DATABASE to accept encoding option. See
README.mb for more details.
For this purpose I have added new column "encoding" to pg_database.
Also pg_attribute and pg_class are changed to catch up the
modification to pg_database. Actually I haved added pg_database_mb.h,
pg_attribute_mb.h and pg_class_mb.h. These are used only when MB is
enabled. The reason having separate files is I couldn't find a way to
use ifdef or whatever in those files. I have to admit it looks
ugly. No way.
* support for PGCLIENTENCODING when issuing COPY command
commands/copy.c modified.
* support for SQL92 syntax "SET NAMES"
See gram.y.
* support for LATIN2-5
* add UNICODE regression test case
* new test suite for MB
New directory test/mb added.
* clean up source files
Basic idea is to have MB's own subdirectory for easier maintenance.
These are include/mb and backend/utils/mb.
calls. Outside a transaction, the backend detects them as buffer
leaks; it sends a NOTICE, and frees them. This sometimes cause a
segmentation fault (at least on Linux). These indexes are initialized
on the first lo_read/lo_write/lo_tell call, and (normally) closed
on a lo_close call. Thus the buffer leaks appear when lo direct
access functions are used, and not with lo_import/lo_export functions
(libpq version calls lo_close before ending the command, and the
backend version uses another path).
The included patches (against recent snapshot, and against 6.3.2)
cause indexes to be closed on transaction end (that is on explicit
'END' statment, or on command termination outside trasaction blocks),
thus preventing the buffer leaks while increasing performance inside
transactions. Some (all?) 'classic' memory leaks are also removed.
I hope it will be ok.
--- Pascal ANDRE, graduated from Ecole Centrale Paris andre@via.ecp.fr
now. Here some tested features, (examples included in the patch):
1.1) Subselects in the having clause 1.2) Double nested subselects
1.3) Subselects used in the where clause and in the having clause
simultaneously 1.4) Union Selects using having 1.5) Indexes
on the base relations are used correctly 1.6) Unallowed Queries
are prevented (e.g. qualifications in the
having clause that belong to the where clause) 1.7) Insert
into as select
2) Queries using the having clause on view relations also work
but there are some restrictions:
2.1) Create View as Select ... Having ...; using base tables in
the select 2.1.1) The Query rewrite system:
2.1.2) Why are only simple queries allowed against a view from 2.1)
? 2.2) Select ... from testview1, testview2, ... having...; 3) Bug
in ExecMergeJoin ??
Regards Stefan
of days --- it was emitting stuff like
REVOKE ALL on 'table' from PUBLIC; GRANT ALL on "table" to
"Public"; neither of which work. While I was at it I
cleaned up a few other things:
* \connect commands are issued only in -z mode. In this way,
reloading a pg_dump script made without -z will generate a simple
database wholly owned by the invoking user, rather than a mishmash
of tables owned by various people but lacking in access rights.
(Analogy: cp versus cp -p.)
* \connect commands are issued just before COPY FROM stdin commands;
without this, reloading a database containing non-world-writable
tables tended to fail because the COPY was not necessarily attempted
as the table owner.
* Redundant \connect commands are suppressed (each one costs a
backend launch, so...).
* Man page updated (-z wasn't ever documented).
The first two items were discussed in a pgsql-hackers thread around
6 May 98 ("An item for the TODO list: pg_dump and multiple table
owners") but no one had bothered to deal with 'em yet.
regards, tom lane
6.3.2 to compile (and run) on my Sparc Solaris 2.5.1 box. Details
below:
pgsql.sparc.patch-template: Adds -D__sparc__ and -D__sun__,
defintions which gcc does define, but Sun's cc does not. :(
pgsql.sparc.patch-makefile: Adds a define so that 'lorder'
is not used, as it is not found on my machine.
Ryan Kirkpatrick
in a more readable form. -- I am submitting the following patches
to the June 6, 1998 snapshot of PostgreSQL. These patches implement
a port of PostgreSQL to SCO UnixWare 7, and updates the Univel port
(UnixWare 2.x). The patched files, and the reason
for the patch are:
File Reason for the patch ---------------
---------------------------------------------------------------
src/backend/port/dynloader/unixware.c src/backend/port/dynloader/unixware.h
src/include/port/unixware.h src/makefiles/Makefile.unixware
src/template/unixware
Created for the UNIXWARE port.
src/include/port/univel.h
Modifed this file to work with the changes made to
s_lock.[ch].
src/backend/storage/buffer/s_lock.c src/include/storage/s_lock.h
Moved the UNIXWARE (and Univel) tas() function from
s_lock.c to s_lock.h. The UnixWare compiler asm
construct is treated as a macro and needs to be in
the s_lock.h file. I also reworked the tas()
function to correct some errors in the code.
src/include/version.h.in
The use of the ## operator with quoted strings in
the VERSION macro caused problems with the UnixWare
C compiler. I removed the ## operators since they
were not needed in this case. The macro expands
into a sequence of quoted strings that will be
concatenated by any ANSI C compiler.
src/config.guess
This script was modified to recognize SCO UnixWare
7.
src/configure src/configure.in
The configure script was modified to recognize SCO
UnixWare 7.
Billy G. Allie
NS32K machine I contributed. In any case, I now have postgresql-6.3
running again on NetBSD/pc532, a NS32532 machine. The following
changes are needed relative to the src directory. (It looks like
support was partially removed when the files were moved from the
src/backend/storage/.... tree to the src/include tree.)
If you need me to get a current development version of postgresql
for this change let me know. Also, let me know if this code needs
updating due to another code movement that deleted the old NS32K
support.
Thank you.
Phil Nelson
GCC, the inner "#if defined(__GNUC__)" can just be omitted in that
architecture's block.
The existing arrangement with an outer "#if defined(__GNUC__)"
doesn't have any obvious benefit, and it encourages missed cases
like this one.
BTW, I'd suggest making the definition of clear_lock for HPUX be
static const slock_t clear_lock = {{-1, -1, -1, -1}};
The extra braces are needed to suppress warnings from gcc, and
declaring it const just seems like good practice.
regards, tom lane
but as I don't have access to a sparc for testing I just did what
I could. I am guessing here, but please apply the following to your
pgsql and let me know what happens. Also, cd to src/storage/buffer
and do 'make s_lock_test' as well.
David Gould
requires manual editing of src/backend/port/getrusage.c, because
its substitute version of getrusage is #if'd out.
There is no good reason for that, because configure won't even
include the file into the Makefile unless the platform hasn't got
getrusage. Furthermore, we only have one working substitute version
of getrusage --- the alleged HPUX syscall-based code doesn't work.
(It causes a coredump because the syscall returns a struct rusage
that's much larger than the stub struct defined in
src/include/rusagestub.h.) The times()-based emulation works fine
on HPUX, however.
I propose, therefore, that getrusage.c should just unconditionally
compile the times-based version, and rely on configure to include
the file only if needed. This will be one less manual configuration
step on all platforms that need this code.
Patch attached.
Tom Lane.
I see someone missed an ancient bit of shell-scripting lore:
on some older shells, if your script's argument list is empty,
then "$@" generates an empty-string word rather than no word
at all. You need to write ${1+"$@"} to get the latter behavior.
(Read your shell man page to see exactly how that works,
but it does the Right Thing on every Bourne shell.)
In particular, pg_dumpall fails when invoked without any switches
on HPUX 9.*, because pg_dump gets an empty-string argument that it
thinks is the name of the database to dump. I expect this bug
also affects some other OSes, but couldn't tell you just which ones.
Patch attached.
The attached patches respond to discussion that was on pgsql-hackers
around the beginning of June (see thread "libpgtcl bug (and symptomatic
treatment)"). The changes are:
1. Remove code in connectDB that throws away the password after making
a connection. This doesn't really add much security IMHO --- a bad guy
with access to your client's address space can likely extract the
password anyway, to say nothing of what he might do directly. And
there's the serious shortcoming that it prevents PQreset() from working
if the database requires a password.
2. Fix coredump problem: fe_sendauth did not guard against being handed
a NULL password pointer. (This is the proximate cause of the coredump-
during-PQreset problem that Magosanyi Arpad complained of last month.)
3. Remove highly questionable "error recovery" logic in libpgtcl's
pg_exec statement.
I believe the consensus of the discussion last month was in favor of
#1 and #3, but I'm just now getting around to making the change.
I realized that #2 was a bug in process of looking at the change.
Attached are diffs (from current cvs sources) to bring libpq.sgml
and libpq.3 up to date.
It appears that at various times in the past, people have made edits to
one or the other of these files but not both. I propagated some changes
from each into the other, but I don't think I caught every
inconsistency. It'd be real nice if the man pages could be
automatically generated from the SGML...
Making PQrequestCancel safe to call in a signal handler turned out to be
much easier than I feared. So here are the diffs.
Some notes:
* I modified the postmaster's packet "iodone" callback interface to allow
the callback routine to return a continue-or-drop-connection return
code; this was necessary to allow the connection to be closed after
receiving a Cancel, rather than proceeding to launch a new backend...
Being a neatnik, I also made the iodone proc have a typechecked
parameter list.
* I deleted all code I could find that had to do with OOB.
* I made some edits to ensure that all signals mentioned in the code
are referred to symbolically not by numbers ("SIGUSR2" not "2").
I think Bruce may have already done at least some of the same edits;
I hope that merging these patches is not too painful.
Used in the generic "CREATE xxx" parsing.
Do some automatic type conversion for inserts from other columns.
Previous trouble with "resjunk" regression test remains for now.
+ Thu Jul 2 20:30:14 CEST 1998
+
+ - Changed new style db name to allow connection types "tcp" and
+ "unix" only
+
+ Tue Jul 7 15:14:14 CEST 1998
+
+ - Fixed some bugs in preproc.y
+ - Set version to 2.3.4
Through some minor changes, I have been able to compile the libpq
client libraries on the Win32 platform. Since the libpq communications
part has been rewritten, this has become much easier. Enclosed is
a patch that will allow at least Microsoft Visual C++ to compile
libpq into both a static and a dynamic library. I will take a look
at porting the psql frontend as well, but I figured it was a good
idea to send in these patches first - so no major changes are done
to the files before it gets applied (if it does).
Regards,
Magnus Hagander
As mentioned around line 1153 in backend/commands/copy.c, the method
of array checking is not perfect.
test=> create table t1 (i text);
test=> insert into t1 values('{\\.}');
INSERT 2645600 1
test=> select * from t1;
i
-----
{\\.}
(2 rows)
test=> copy t1 to '/tmp/aaa';
test=> copy t1 from '/tmp/aaa';
ERROR: CopyReadAttribute - end of record marker corrupted
Copy cannot read data produced by itself!
I have implemented a framework of encoding translation between the
backend and the frontend. Also I have added a new variable setting
command:
SET CLIENT_ENCODING TO 'encoding';
Other features include:
Latin1 support more 8 bit cleaness
See doc/README.mb for more details. Note that the pacthes are
against May 30 snapshot.
Tatsuo Ishii
This incorporates all the precedeing patches and emailed suggestions
and the results of the performance testing I posted last week. I
would like to get this tested on as many platforms as possible so
I can verify it went in correctly (as opposed to the horrorshow
last time I sent in a patch).
Once this is confirmed, I will make a tarball of files that can be
dropped into a 6.3.2 source tree as a few people have asked for
this in 6.3.2 as well.
David Gould
syntax that can be read back in with psql. I did this by adding
a
"-c" switch that controls moving the CONTSTRAINT statements inside
the CREATE TABLE statements and adding () around the CHECK arguments.
Here's diffs against the 6.3.2 version of pg_dump.c.
ccb
Attached to the mail is locale-patch.tar.gz. In the archive
there are:
file README.locale
short description
directory src/test/locale
test suite; currently only koi8-r tests, but the suite can be
easily extended
file locale.patch
the very patch; to apply: patch < locale.patch; should be applied
to postgres-6.3.2 (at least I created it with 6.3.2 without any
additional
patches)
Files touched by the patch: src/include/utils/builtins.h
src/backend/utils/adt/char.c src/backend/utils/adt/varchar.c
src/backend/utils/adt/varlena.c
Oleg
pg_notifies statement is eliminated, and callbacks defined by
pg_listen are instead invoked automatically from the Tcl idle loop
whenever a NOTIFY message is received.
I have done only cursory testing, so there may be problems still
lurking (particularly on non-Unix machines?). But it seems to
work.
Patch is against today's cvs sources. Note that this will not work
with the 6.3.2 release since it depends on the new libpq.
The diffs are a bit large so I've gzipped them. A patch to update
libpgtcl.sgml is included too.
regards, tom lane
configuration system. The idea is to make the configure arguments
that specify compilers to be compatible with the other --with
options. The main point, though, is that the c++ support is on by
default, but can easily be disabled by the --without-CXX option
for those few(?) that don't want it.
Brook Milligan
have > 20000 users and each (potentially) needs a separate database
which is > only accessible to them. Rather than having 20000 lines
in pg_hba.conf, > I've patched Postgres so that the special token
"sameuser" in the > database field of pg_hba.conf allows access
only to the username which > is connecting.
things as well:
* Computes and saves a cancel key for each backend. * fflush
before forking, to eliminate double-buffering problems
between postmaster and backends.
Other cleanups.
Tom Lane
src/test/regess/sql/junkfilter.sql -- SQL for
regression test src/test/regess/expected/junkfilter.out --
Expected output SQL for regression test
David Hartwig
Bug fixes:
PreparedStatement.setObject didn't handle short's
ResultSet.getDate() now handles null dates (returns null rather
than a NullPointerException)
ResultSetMetaData.getPrecision() now returns 0 for VARCHAR
New features:
Field now caches the typename->oid in a Hashtable to speed things
up. It removes the need for some unnecessary queries to the
backend.
PreparedStatement.toString() now returns the sql statement that
it will send to the backend. Before it did nothing.
DatabaseMetaData.getTypeInfo() now does something.
Bug fixes:
PreparedStatement.setObject didn't handle short's
ResultSet.getDate() now handles null dates (returns null rather
than a NullPointerException)
ResultSetMetaData.getPrecision() now returns 0 for VARCHAR
New features:
Field now caches the typename->oid in a Hashtable to speed things
up. It removes the need for some unnecessary queries to the
backend.
PreparedStatement.toString() now returns the sql statement that
it will send to the backend. Before it did nothing.
DatabaseMetaData.getTypeInfo() now does something.
+ Wed Jun 3 13:38:57 CEST 1998
+
+ - Made sqlca struct compatible with other systems.
+ - Give back a warning in case of truncation
+ - Changed the handling of OptimizableStmt since the old one broke
+ CREATE RULE
+ - Set library version to 2.3
+ - Set version to 2.3.3
and vica versa for the next operation.
This is reputed to fix inline math optimization troubles in glibc-2.0.x.
Regression tests still pass on libc/i686 machine. Patch suggested by Matt.
Change ordering of IS_BUILTIN_TYPE() macro to optimize lookup.
Make CASH type _not_ equivalent to INT4.
CASH is passed by reference rather than passed by value.
Currently force the type to match the _first_ select in the union.
Move oper_select_candidate() from parse_func.c to parse_oper.c.
Throw error inside of oper_inexact() if no match for binary operators.
Check more carefully that types can be coerced
even if there is only one candidate operator in oper_inexact().
Fix up error messages for more uniform look.
Remove unused code.
Fix up comments.
allowed to be used for alternate database locations.
Probably best to default to not allowed, as now, since there are security
and integrity issues which should be considered carefully before
opening things up.
Will update docs to discuss this issue.
Add additional tests in strings for conversions of the "name" data type.
Test SQL92 string functions such as SUBSTRING() and POSITION().
Fix geometry tests to reflect code fixed by Gautam.
Update error messages.
+Wed May 20 10:46:48 CEST 1998
+
+ - Fixed handling of preprocessor directives and variable
+ initialization.
+ - Added enum datatype.
- Set version to 2.3.2
Everything (except of course random) passes on my netbsd box except int2,
int4, oidint2, and oidint4; all fail because of error message differences.
Below are some patches to the expectations to correct the problem by creating
*-NetBSD.out files.
+
+ Wed May 6 16:09:45 CEST 1998
+
+ - Some more cleanups in the library.
+
+ Thu May 7 12:34:28 CEST 1998
+
+ - Made CONNECT and DISCONNECT statement more SQL3 compliant.
+ - Changed the API for the ECPGconnect function to be able to handle
+ hostnames and ports
+
+ Fri May 8 13:54:45 CEST 1998
+ - More changes to the parser. The connect statement now allows
+ ORACLE style logins.
+ - db-name is accepted in two ways:
+ - <dbname>[@<server>][:<port>]
+ - esql:postgresql://<server>[:<port>][/<dbname>]
+
+ Mon May 11 10:28:37 CEST 1998
+
+ - Added '? options' to connect call.
+ - Also allow USING as keyword for the password
+
+ Thu May 14 15:09:58 CEST 1998
+
+ - Changed preproc.y and pgc.l according to the parser changes in the
+ backend.
+
+ Fri May 15 09:55:21 CEST 1998
+
+ - Added connection_name handling
+
+
+ Mon May 18 10:33:58 CEST 1998
+
+ - Fixed some more bugs
+ - Set version to 2.3.1
+ - Set library version to 2.2
psql in Postgres 6.3.2. Both of these problems were complained of
recently in pgsql-questions:
1. In the right circumstances, psql.c will fail to compile due to
trying
to include a nonexistent <history.h>. (Thread "Compile-time
error" around 17 Apr 98.) 2. In other circumstances, psql will
compile but does not provide
command history capability, even though the underlying readline
library supports it. (Various threads, most recently "query
repetition in psql" around 29 Apr.)
Tom Lane
Check for and properly ignore FOREIGN KEY column constraints
(already had fixed same for table constraints).
Define USER as synonym for CURRENT_USER per SQL92 specs
(no longer allowed as bare column name, sorry).
Re-enable HAVING clause but no fixes elsewhere yet.
Other stuff from today's update of gram.y...
Check for and properly ignore FOREIGN KEY column constraints
(already had fixed same for table constraints).
Define USER as synonym for CURRENT_USER per SQL92 specs
(no longer allowed as bare column name, sorry).
Re-enable HAVING clause but no fixes elsewhere yet.
Make "char" type a synonum for "char(1)" (actually implemented as bpchar).
Compress/compact row-style subselect and operator definitions
(cut out ~140 lines of code with no change in functionality).
Save string type if specified for DEFAULT clause handling.
Enough for now...
1. Rewritten libpq to allow asynchronous clients.
2. Implemented client side of cancel protocol in library,
and patched psql.c to send a cancel request upon SIGINT. The
backend doesn't notice it yet :-(
3. Implemented 'Z' protocol message addition and renaming of
copy in/out start messages. These are implemented conditionally,
ie, the client protocol version is checked; so the code should
still work with 1.0 clients.
4. Revised protocol and libpq sgml documents (don't have an SGML
compiler, though, so there may be some markup glitches here).
What remains to be done:
1. Implement addition of atttypmod field to RowDescriptor messages.
The client-side code is there but ifdef'd out. I have no idea
what to change on the backend side. The field should be sent
only if protocol >= 2.0, of course.
2. Implement backend response to cancel requests received as OOB
messages. (This prolly need not be conditional on protocol
version; just do it if you get SIGURG.)
3. Update libpq.3. (I'm hoping this can be generated mechanically
from libpq.sgml... if not, will do it by hand.) Is there any
other doco to fix?
4. Update non-libpq interfaces as necessary. I patched libpgtcl
so that it would compile, but haven't tested it. Dunno what
needs to be done with the other interfaces.
Have at it!
Tom Lane
Tue Apr 28 14:48:41 CEST 1998
- Put operator "->" back into parser. Note that :foo->bar means the
C term, but :foo ->bar means the operator "->".
Tue Apr 28 15:49:07 CEST 1998
- Added exec sql disconnect command.
- Allow varchar in C to be written in uppercase too.
- Added whenever option "do break;"
Wed Apr 29 09:17:53 CEST 1998
- Corrected parsing of C comments.
- Also allow C++ style comments.
- Make sure not found is only checked after commands that could
return it.
- Added error codes, see ecpgerror.h for details.
- Added "exec sql <TransactionStmt> release" as disconnect statement
for compatibility issues.
Thu Apr 30 10:42:10 CEST 1998
- Added a -t option to disable automatic transaction start.
- Added sqlerrd[] to sqlca struct.
- Give back number of tuples affect in sqlca.sqlerrd[2].
Thu Apr 30 13:36:02 CEST 1998
- Make the return code different in case of different errors.
Wed May 6 11:42:48 CEST 1998
- Free memory if possible
- Some bugfixes for bugs I found while changing the memory
allocation code
- Now able to fill complete array with one call (see test1.pgc for
an example)
- Set version to 2.3.0
- Set library version to 2.1
Attached patch will add a version() function to Postges, e.g.
template1=> select version();
version
------------------------------------------------------------
PostgreSQL 6.3.2 on i586-pc-linux-gnu, compiled by gcc 2.8.1
(1 row)
If PQfn() receives NOTICEs from the backend, it fails because there is no
provision to deal with them.
This patch (supplied by Anders Hammarquist <iko@netg.se> to me as Debian
maintainer of postgresql) cures the problem:
Hi, here are patches I promised (against 6.3.2):
* character_length(), position(), substring() are now aware of
multi-byte characters
* add octet_length()
* add --with-mb option to configure
* new regression tests for EUC_KR
(contributed by "Soonmyung. Hong" <hong@lunaris.hanmesoft.co.kr>)
* add some test cases to the EUC_JP regression test
* fix problem in regress/regress.sh in case of System V
* fix toupper(), tolower() to handle 8bit chars
note that:
o patches for both configure.in and configure are
included. maybe the one for configure is not necessary.
o pg_proc.h was modified to add octet_length(). I used OIDs
(1374-1379) for that. Please let me know if these numbers are not
appropriate.
HP-UX (all versions) requires shared libraries to have execute
permission, and really needs them to be exactly mode 555 for
performance reasons. The standard configure/install procedure
installs libpq.sl as mode 644, which DOES NOT WORK.
The attached patch modifies the makefiles to distinguish
INSTL_LIB_OPTS (install mode for ordinary libraries) from
INSTL_SHLIB_OPTS (mode for shared libs), and adds a test
to configure to set INSTL_SHLIB_OPTS="-m 555" when on HP-UX.
Ok, I have finally gotten all of the defines for Dec/Alpha and
Linux/Alpha sorted out as Marc asked. There is no longer any need for
'-Dalpha' or '-Dlinuxalpha' in either the Dec/Alpha or the Linux/Alpha
template files (./src/template/{alpha,linuxalpha}). I have replaced every
instance of 'alpha' or '__alpha__' with '__alpha', as that appears to be
the common symbol between C compilers on both operating systems (RH4.2 &
DecUnix 4.0b) for alpha.
1. Removes the unnecessary "#define AbcRegProcedure 123"'s from
pg_proc.h.
2. Changes those #defines to use the names already defined in
fmgr.h.
3. Forces the make of fmgr.h in backend/Makefile instead of having
it
made as a dependency in access/common/Makefile *hack*hack*hack*
4. Rearranged the #includes to a less helter-skelter arrangement,
also
changing <file.h> to "file.h" to signify a non-system header.
5. Removed "pg_proc.h" from files where its only purpose was for
the
#defines removed in item #1.
6. Added "fmgr.h" to each file changed for completeness sake.
Turns out that #6 was not necessary for some files because fmgr.h
was being included in a roundabout way SIX levels deep by the first
include.
"access/genam.h"
->"access/relscan.h"
->"utils/rel.h"
->"access/strat.h"
->"access/skey.h"
->"fmgr.h"
So adding fmgr.h really didn't add anything to the compile, hopefully
just made it clearer to the programmer.
S Darren.
+ Thu Apr 23 09:27:16 CEST 1998
+
+ - Also allow call in whenever statement with the same functionality
+ as do.
+
+ Thu Apr 23 12:29:28 CEST 1998
+
+ - Also rewrote variable declaration part. It is now possible to
+ declare more than one variable per line.
+ - Set version to 2.1.0
+
+ Fri Apr 24 13:50:15 CEST 1998
+
+ - Fixed some bugs.
+ - Set version to 2.1.1
Here's a fix for a tiny memory leak in PQsetdb/PQfinish.
(Analysis of a running program indicates there are several others, but
this is the only obvious one I saw in the code).
This fixes a problem in ResultSet.getDate() when the column is NULL
(reported by Vincent Partington <Vincent.Partington@nmg.nl>)
And fixes a problem with Field's (ResultSet.getObject() was proving to be
slow as it repetedly send queries for oid -> name mapping - fixed by
creating a cache. (reported by Mario Ellebrecht <ellebrec@nads.de>)
Here is a pair of patches that (I hope) finish the configuration
issues with tcl/tk and make the recognition of the two packages
completely parallel in organization. This should make future changes
easier to maintain.
Hope to see this in 6.2.2.
and TIMEZONE_MINUTE but don't introduce until v6.4.
Fix SET TIMEZONE LOCAL to pass null pointer
rather than older "default" string.
Fix handling of NULL pointer returns from FOREIGN KEY clauses
which are currently ignored.
Allow START as a table/column name.
Attached you'll find a (big) patch that fixes make dep and make
depend in all Makefiles where I found it to be appropriate.
It also removes the dependency in Makefile.global for NAMEDATALEN
and OIDNAMELEN by making backend/catalog/genbki.sh and bin/initdb/initdb.sh
a little smarter.
This no longer requires initdb.sh that is turned into initdb with
a sed script when installing Postgres, hence initdb.sh should be
renamed to initdb (after the patch has been applied :-) )
This patch is against the 6.3 sources, as it took a while to
complete.
Please review and apply,
Cheers,
Jeroen van Vianen
probleme number 1 :
- configure can find the library readline , but don't
find the header file . so in this case we don't use lib readline
.
probleme number 2 :
- when you have postgres 6.2.1 and readline installed
with the same prefix( and generally all your software ) . you
can compile the version 6.3 . I use this prefix , when configure
ask me for "Additional directories to search for include files"
.
( because there a conflict in the header when you
compile psql.c ) In this case, you must permut the sequence of
directive -I .
Erwan MAS
the configuration of v6.3.1. I have replaced the queries for
include/lib directories with --with configuration options. I have
also included a list of potential tcl/tk include directories directly
in the CPPFLAGS variable. As new versions are needed, these should
be added to the list in reverse numerical order (libraries are in
a separate list near the end). This greatly simplifies the later
checks if --with-tcl is set. I hope this solution works for
everyone.
I also added a check to disable the perl support if postgres was
not already installed (as per the instructions in the directory).
By the way, why must there be an installed pgsql to compile perl
support? This seems odd, at best.
Finally, I changed the Makefile in the libpgtcl interface to place
the shared libraries at the end of the list of files, not at the
beginning. With NetBSD at least, libraries are linked in order,
so the original sequence does not work.
Brook Milligan
However somebody else also applied a patch to the same part of
configure to fix a different problem. So part of my patch was not
applied or got reversed or ... whatever.
The attached patch will restore configure --with-tcl to working
order and should remove a lot of the messages complaining about
tcl not working.
Alvin
After applying the following patch there remain two
probable buffer overruns detected by Electric Fence during
the regression test.
I'll try find out what causes the remain two ones.
This patch also corrects a typo in smgr.c.
3) Add "#include "config.h" to src/interfaces/ecpg/preproc/pgc.l
to correct "strings.h not found". config.h has the proper define to
make this work and should probably be near the top of pgc.l before
the first include.
2) Add "#define gettimeofday(a,b) gettimeofday(a) to src/include/config.h
On the 88k SVR4, gettimeofday only has one argument. This is
checked for in a few other packages by configure, so there should
be some examples of the configure test out there.
was detected by Electric Fence and triggered by statements like:
SELECT * into table t from pg_database;
The system would crash on a memmove call in DataFile() with arguments
like this:
memmove(0x0, 0x0, 0);
Maurice Gittens
was a 2000 character buffer allocated for results, and the files
you refer to produce a 2765 byte column called formsource. This
should not have worked with any version of libpgtcl.
Nevertheless, the limit is an artificial one, since there is no
need to use this intermediate buffer where it is being used and
abused.
Randy Kunkee <kunkee@pluto.ops.NeoSoft.com>
1. Remove the char2, char4, char8 and char16 types from postgresql
2. Change references of char16 to name in the regression tests.
3. Rename the char16.sql regression test to name.sql. 4. Modify
the regression test scripts and outputs to match up.
Might require new regression.{SYSTEM} files...
Darren King
access overrun. For the sake of doing things properly here is a
patch which fixes it.
This patch is for the file backend/commands/sequence.c.
Maurice Gittens
yyerror ones from bison. It also includes a few 'enhancements' to
the C programming style (which are, of course, personal).
The other patch removes the compilation of backend/lib/qsort.c, as
qsort() is a standard function in stdlib.h and can be used any
where else (and it is). It was only used in
backend/optimizer/geqo/geqo_pool.c, backend/optimizer/path/predmig.c,
and backend/storage/page/bufpage.c
> > Some or all of these changes might not be appropriate for v6.3,
since we > > are in beta testing and since they do not affect the
current functionality. > > For those cases, how about submitting
patches based on the final v6.3 > > release?
There's more to come. Please review these patches. I ran the
regression tests and they only failed where this was expected
(random, geo, etc).
Cheers,
Jeroen
sequential scans! (I think it will also work with hash, index, etc
but I did not check it out! I made some High level changes which
should work for all access methods, but maybe I'm wrong. Please
let me know.)
Now it is possible to make queries like:
select s.sname, max(p.pid), min(p.pid) from part p, supplier s
where s.sid=p.sid group by s.sname having max(pid)=6 and min(pid)=1
or avg(pid)=4;
Having does not work yet for queries that contain a subselect
statement in the Having clause, I'll try to fix this in the next
days.
If there are some bugs, please let me know, I'll start to read the
mailinglists now!
Now here is the patch against the original 6.3 version (no snapshot!!):
Stefan
a dumpall. This has been happening when a second \connect is
encountered.
The faulty code was in fe-connect.c, where the memory for the user
password was freed, but the pointer itself was not set to NULL.
Later, the memory was reused and the password appeared not to be
empty, so that an attempt was made to reference it.
Oliver Elphick
1) DatabaseMetaData.getPrimaryKeys() would fail saying that there
is no
table t.
2) PreparedStatement.getObject() was missing some break statements,
which
was causing updates not to work with JBuilder (supplied by Aaron
Dunlop).
jdbc fixes from Peter.
manager to not try to split files in 2 gig chunks. It will just
try to get another block.
If applied, everything is just as before. But if LET_OS_MANAGE_FILESIZE
is defined, the chaining disappears and the file just keeps on
going, and going, and going, til the OS barfs.
Darren King
--------------------------------------------------------------------------
Test Case: ----------
--------------------------------------------------------------------------
Solution: --------- Add this to the libpq and libpq++ Makefiles
to build shared libs:
Mike Ferrara
extern char *sys_errlist[]; #define strerror(A) (sys_errlist[(A)])
#endif /* sunos4
*/
is picked up by Solaris when the above is intended only for SunOS.
Fix Solaris. Albert Chin-A-Young
to main.c are only to add some extra includes to support some code
that's suddenly being used.
The #define ASSEMBLER is to prevent most of the code of sys/proc.h
from being included, as it ends up conflicting with some of the
postgresql definitions. This may or may not work on other versions
of Digital Unix.
Get alpha working. Yea. Dwayne Bailey
> > characters in them. Dumping and reloading using pg_dumpall >
> doesn't work with this and dumping the entire array and > > then
trying to parse it is hopeless.
Doug Gibson
Make "TABLE" optional in "LOCK TABLE" command
and "... INTO TABLE..." clause.
Explicitly parse CREATE SEQUENCE options to allow a negative integer
as an argument; this is an artifact of unary minus handling in scan.l.
Add "PASSWORD" as an allowed column identifier.
These fixes will require a "make clean install" but not a dump/reload.
a while back I posted a patch for pg_ident, the patch worked but I didn't
diagnose the problem properly.
on my compiler(gcc2.7.2) this compiles with no errors...
char buf[1000]; if(buf != '\0') {
...but it doesn't compare '\0' with the first char of buf.
There is an error in the configure script when using
--with-pgport= that will cause the compiled version of
PostgreSQL to no longer allow connections to the
new port and to treat shared memory improperly.
What happens is that if the port is changed, the configure
script defines DEF_PGPORT as "", which atoi() will return
as 0, which makes the IPC_KEY value 0. This then causes
semaphores to be allocated, but never released. Postgres
eventually returns from semget() with
"no space left on device". The source of this error could
easily be overlooked in version 6.3 since it is possible
to connect via UNIX domain sockets, and having DEF_PGPORT
defined as "0" would not be noticed until TCP was used.
The following patch is to src/interfaces/libpq of postgresql-6.3.
The purpose of the patch is to make the initialization of
const char *pgresStatus[] match the ExecStatusType enum.
6.3 postmaster is supposed to work with pre 6.3 protocol. This is true
for little endian architecture servers. But for big endian machines
such as Sparc the backward compatibility function do not work.
Attached are patches to fix the problem.
For substr() and substring() on the text data type, the relevant code is in
varlena.c. You are right, there is a problem. I have a patch which I will
apply to the source tree soon. The copy enclosed below probably does not
preserve tabs correctly so cannot be applied directly; the relevant change
is simply changing the ">=" to ">"...
It is my hope that the following "patches" to libpgtcl get included
in the next release.
See the update to the README file to get a full description of the changes.
This version of libpgtcl is completely interpreter-safe, implements the
database connection handle as a channel (no events yet, but will make it
a lot easier to do fileevents on it in the future), and supports the SQL
"copy table to stdout" and "copy table from stdin" commands, with the
I/O being from and to the connection handle. The connection and result
handles are formatted in a way to make access to the tables more efficient.
Included are patches intended for allowing PostgreSQL to handle
multi-byte charachter sets such as EUC(Extende Unix Code), Unicode and
Mule internal code. With the MB patch you can use multi-byte character
sets in regexp and LIKE. The encoding system chosen is determined at
the compile time.
To enable the MB extension, you need to define a variable "MB" in
Makefile.global or in Makefile.custom. For further information please
take a look at README.mb under doc directory.
(Note that unlike "jp patch" I do not use modified GNU regexp any
more. I changed Henry Spencer's regexp coming with PostgreSQL.)
Included are patches intended for allowing PostgreSQL to handle
multi-byte charachter sets such as EUC(Extende Unix Code), Unicode and
Mule internal code. With the MB patch you can use multi-byte character
sets in regexp and LIKE. The encoding system chosen is determined at
the compile time.
To enable the MB extension, you need to define a variable "MB" in
Makefile.global or in Makefile.custom. For further information please
take a look at README.mb under doc directory.
(Note that unlike "jp patch" I do not use modified GNU regexp any
more. I changed Henry Spencer's regexp coming with PostgreSQL.)
Ok, this fixes three things:
1. It seems (from tests submitted by two people with JBuilder) that
JBuilder expects a responce from ResultSetMetaData.getPrecision() &
getScale() when used on non numeric types. This patch makes these
methods return 0, instead of throwing an exception.
2. Fixes a small bug where getting the postgresql type name returns null.
3. Fixes a problem with ResultSet.getObject() where getting it's string
value returns null if you case the object as (PGobject), but returns
the value if you case it as it's self.
Patch1:
Postgres thinks dist_pl (dist of a point to a line) is expecting a box (603)
for the right arg, but it really should be a line (628).
Otherwise the left & right args match those of dist_pb (dist of a point to a
box) two lines further down.
Patch2:
Anyways, these two functions take a path (602) whereas in pg_proc.h they are
listed as taking a lseg (601).
1. Make 'all' works without complaint. Don't have to add the .exp
files to the files list. They are made automagically when
making the respective shared lib file.
Only port that actually uses EXPSUFF (from makefiles/Makefile.*)
is Aix, so if this breaks anybody else, let me know, asap.
2. Make 'clean' actually cleans up correctly. Previously, it would
leave the .o files in C-code directory.
3. Changed references to reflect new location of .c files.
4. Added DELETE statements to complex.source so that it tidies up
when done. Previously, it would leave things in pg_amop,
pg_amproc and pg_opclass. Only possible to do this with the
new SUBSELECT code in 6.3. Nice work, fellas...
Not deleting the index entries would cause a non-fatal error if
complex.sql was run again on the same database. Much tidier now.
5. Corrected the README. obj directory hasn't existed since Bryan
redid the make way back when. Also changed the snipet from psql
to match the current version. POSTGRES95?!? I don't think so. :)
The following patches will allow postgreSQL 6.3 to compile and run on a
UNIXWARE 2.1.2 system with the native C compiler with the following library
change:
The alloca function must be copied from the libucb.a archive and added
to the libgen.a archive.
Also, the GNU flex program is needed to successfully build postgreSQL.
Seem to remember someone posting to one of the lists a while back
that the tutorial code wouldn't compile and/or run. Found four
problems with it that will let it run.
1. Tutorial makefile had a recursive use of DLOBJS.
2. Some tutorial needed semi-colons added to many statements.
3. Complex tutorial didn't clean up after itself.
4. Advanced had a time-travel example. Commented it out and
put a line pointing the user to contrib/spi/README.
Two incorrect printf formats in parser/parse_type.c. Prolly done
by me a long time ago when I cleaned up int's and Oid's...
Format flag is really just %u, not %ud. Harmless, but results in
"type id lookup of 25d failed" instead of only "...25 failed"
This patch will...
1. Remove the "-Wall" option from the ecpg/lib and ecpg/preproc Makefile.
2. Remove the addition of $(SRCDIR)/include and-or $(SRCDIR)/backend from
ecpg/lib, ecpg/preproc, libpq and utils Makefiles. Already in CFLAGS...
3. Set MK_NO_LORDER and RANLIB in Makefile.aix to avoid a couple of extra
steps taken care of by the 'ld' command anyways.
I thought it would be a good idea to ensure that the new view
permission model will not get broken by subsequent
fixes/changes. So I wrote a little regression test for it.
There is an ugly thing in this regression test. It creates
temporary a test user that is required for the tests. The
user is removed at the end of the test, but if sometimes the
regression suite is aborted or crashes exactly here, the test
user will lay around in the pg_shadow. Don't have a clue how
to get around.
return, not a slot returned from access method (they have
different TupleDesc and MergeJoin node was broken).
nodeIndexscan.c: index_markpos()/index_restrpos() call index-specific
mark/restr funcs and are in use now (instead of
IndexScanMarkPosition()/ExecIndexRestrPos()).
seems that my last post didn't make it through. That's good
since the diff itself didn't covered the renaming of
pg_user.h to pg_shadow.h and it's new content.
Here it's again. The complete regression test passwd with
only some float diffs. createuser and destroyuser work.
pg_shadow cannot be read by ordinary user.
+
+ - use char[] as string not as array of bytes that is integers
+
+ Sun Feb 22 16:37:36 CET 1998
+
+ - use long for all size variables
+ - added execute immediate statement
+
+ Sun Feb 22 20:41:32 CET 1998
+
+ - use varcharsize = 1 for all simple types, 0 means pointer, > 1
+ means array if type is char resp. unsigned char
+
+ Thu Feb 24 12:26:12 CET 1998
+
+ - allow 'go to' in whenever statement as well as 'goto'
+ - new argument 'stop' for whenever statement
From: Michael Meskes <meskes@topsystem.de>
What it does:
It solves stupid problem with cyrillic charsets IP-based on-fly recoding.
take a look at /data/charset.conf for details.
You can use any tables for any charset.
Tables are from Russian Apache project.
Tables in this patch contains also Ukrainian characters.
Then run ./configure --enable-recode
Ok. I have decided to use:
#if defined(sun) && if defined(sparc) && !defined(__svr4)
instead of defined(sunos4). interfaces/libpq/libpq-fe.h and
include/c.h have been modified(see included patches).
Another porblems I have found are:
o SunOS lacks strtoul(). to fix this I stole strtoul.c from FreeBSD
and place it under backend/port. necessary modifications have been
also made to backend/port/Makefile.in, include/config.h.in and
configure.in (see included patches).
So if the relname is given to acldefault() in
utils/adt/acl.c, it can do a IsSystemRelationName() on it and
return ACL_RD instead of ACL_WORLD_DEFAULT.
The diff looks so simple and easy. But to find it wasn't fun.
It must have been there for a long time. What happened:
When a tuple in one of some central catalogs was updated, the
referenced relation got flushed, so it would be reopened on
the next access (to reflect new triggers, rules and table
structure changes into the relation cache).
Some data (the tupleDescriptor e.g.) is used in the system
cache too. So when a relation is subject to the system cache,
this must know too that a cached system relation got flushed
because the tupleDesc data gets freed during the flush!
For the GRANT/REVOKE on pg_class it was slightly different.
There is some local data in inval.c that gets initialized on
the first invalidation of a tuple in some central catalogs.
This needs a SysCache lookup in pg_class. But when the first
of all commands is a GRANT on pg_class, exactly the needed
tuple is the one actually invalidated. So I added little code
snippets that the initialization of the local variables in
inval.c will already happen during InitPostgres().
Enclosed is the regression.diffs file from running the Feb 21st
snapshot regression tests for inclusion in src/test/regression
as regression.Aix41. Appears to be standard differences to me,
error messages, fp accuracy and times off by an hour due to PST
vs PDT.
#define TAPETEMP "pg_btsortXXXXXX"
to:
#define TAPETEMP "pg_btsortXXXXXXX"
For some reason, under FreeBSD, it appears that the mktemp() value needs the
extra 'X' to improve/ensure uniqueness
below is the patch to have views to override the permission
checks for the accessed tables. Now we can do the following:
CREATE VIEW db_user AS SELECT
usename,
usesysid,
usecreatedb,
usetrace,
usecatupd,
'**********'::text as passwd,
valuntil
FROM pg_user;
REVOKE ALL ON pg_user FROM public;
REVOKE ALL ON db_user FROM public;
GRANT SELECT ON db_user TO public;
any other, example program.
I have tracked this down to a call to PQfinish() in ECPGfinish()
that occurs before any connection is established.
From: Keith Parks <emkxp01@mtcc.demon.co.uk>
whatsoever. The patch is not a solution, because configure is generated
from configure.in, and I don't know how to patch it to get a working
'configure'.
From: "Pedro J. Lobo" <pjlobo@euitt.upm.es>
dgux 5.4R4.11
Missing port-protos.h (not needed, I think). Wants dld.h. Should
really use the system dl stuff (like i386_solaris). Needs to include
<netinet/in.h> before <arpa/inet.h>. Here are some patches...
compiler define that should have been enabled, but was not due to
different naming conventions for Linux/Alpha. Attached is the patch he
sent me, that I have not had a chance to test yet.
From: Ryan Kirkpatrick <rkirkpat@nag.cs.colorado.edu>
The file 'backend/utils/adt/arrayfuncs.c' won't compile with the
February 14th snapshot, because of an inconsistency between the
declaration and implementation of ReadArrayStr(). As far as I can
tell, the predeclaration is wrong. I assume this is what was meant:
just a little correction in the pltcl_guide.nr.
Sometimes I changed the name of tuple arguments to numbers
like the other args are. Otherwise it wasn't possible to
create a function as
CREATE FUNCTION f (EMP, EMP) ... LANGUAGE 'pltcl';
The arguments are now accessed in the function as
$1(name) vs. $2(name)
Only occurrs in
src/include/storage/s_lock.h:#if defined(__AIX)
src/include/utils/dt.h:#if defined(__AIX)
src/include/utils/nabstime.h:#if defined(__AIX)
Simply delete one underscore, only occurs once per file, so no patch.
Someone changed the parser to build a TypeName node on CREATE
FUNCTION in any case. As a side effect, ALL! functions
created got the proretset attribute to true. Thus for a
SELECT the parser wrapped an Iter node around the Expr and
since singleton functions set isDone the Iter returns no
tuple up.
Apart from this Makefile hack, all I've done is to make dynamically
loaded code modules fail properly (as was already done for __mips__,
although I think this is too loose: I believe NetBSD for the pmax can
do dynamic linking), and to add test-and-set lock handling. As Bruce
suggested, this is done in a maximally efficient inlined way: I was
not aware that this code was so important, speed-wise.
of some global variables to support subselects and calls union_planner().
Calls to SS_replace_correlation_vars() and SS_process_sublinks() in
query_planner() before planning.
Get rid of #ifdef INDEXSCAN_PATCH in createplan.c.
ExecReScan for nodeAgg, nodeHash, nodeHashjoin, nodeNestloop and nodeResult.
Fixed ExecReScan for nodeMaterial.
Get rid of #ifdef INDEXSCAN_PATCH.
Get rid of ExecMarkPos and ExecRestrPos in nodeNestloop.
A few minutes ago I sent down the PL/Tcl directory to this
list. Look at it and reuse anything that might help to build
PL/perl. I really hope that PL/perl and PL/Tcl appear in the
6.3 distribution. I'll do whatever I can to make this happen.
A few minutes ago I sent down the PL/Tcl directory to this
list. Look at it and reuse anything that might help to build
PL/perl. I really hope that PL/perl and PL/Tcl appear in the
6.3 distribution. I'll do whatever I can to make this happen.
A few minutes ago I sent down the PL/Tcl directory to this
list. Look at it and reuse anything that might help to build
PL/perl. I really hope that PL/perl and PL/Tcl appear in the
6.3 distribution. I'll do whatever I can to make this happen.
Use explicit tokens to decode CREATE TRIGGER clauses.
Allow ROW and STATEMENT as column identifiers.
Fix CAST syntax to require parens per SQL92 spec.
Define TypeId to allow correct translation of type names in CREATE FUNCTION
and other statements. Need to do this without looking up defined type
names because CREATE FUNCTION can specify undefined (new) types.
Define UserId to complete removal of "Id" generic entity.
Define xlateSqlFunc() to convert SQL92 CHARACTER_LENGTH() and CHAR_LENGTH()
functions to calls to length().
Define func_name parser entity for contexts requiring a function name.
Have xlateSqlType() translate "float" to "float8".
This patch fixes the following:
* Fixes minor bug found in DatabaseMetaData.getTables() where it doesn't
handle default table types.
* It now reports an error if the client opens a database using
properties, and either the user or password properties are missing. This
should make the recent problem with Servlets easier to find.
* Commented out obsolete property in Driver.getPropertyInfo()
Well this is not really a patch. But I mananged to get Linus' old Postgres95
precompiler to compile and work with PostgreSQL. The next step would be to
collect bug/missing feature reports and to put it into the distribution so
that it is made with the standard make procedure.
Warning! So far it is not tested much and it does not install correctly. But
I was able to create a small binary with it.
select from a table with attrs (a int, b char(20))
crashed in bpcharout() (palloc of -1 bytes). But a table
with attrs (a int, b varchar(20)) worked.
From: Jan Wieck <jwieck@debis.com>
surrounded by parentheses (but not all are meaningful).
Remove unused keywords ACL, APPEND, MERGE.
Requires a "make clean" to recompile all code since keyword numeric
assignments have changed with keyword removal.
Define functions and operators for closest point to lseg on box,
to line on lseg, to lseg on lseg.
Define function and operator for length of lseg.
Change length operator from '??' to '@-@'
(currently defined for path and lseg).
Define close_ls(), close_lseg(), lseg_length().
Write real code for close_sb(), close_pb(), inter_sb(), inter_lb().
Repair lseg_perp() which determines if two lsegs are perpendicular.
Repair lseg_dt() distance between two lsegs.
Note: close_sl() is clearly broken but will repair later
(calculating point on lseg rather than point on line).
[This is a repost - it supercedes the previous one. It fixes the patch so
it doesn't bread aix port, plus there's a file missing out of the
original post because difforig doesn't pick up new files. It's now
attached. peter]
This patch brings the JDBC driver up to the current protocol spec.
Basically, the backend now tells the driver what authentication scheme to
use.
The patch also fixes a performance problem with large objects. In the
buffer manager, each fastpath call was sending multiple Notifications to
the backend (sometimes more data in the form of notifications were being
sent than blob data!).
if an operating specific expected file exists, use that for the comparison.
This allows for "legit" differences between results, like the "Result too
large" message vs "Math result not representable" ...
Also, have the failed diffs get output to regression.diffs so that its easy to
view those tests that failed
I haven't had final confirmation from Peter yet, but the attached patch
needs to be applied for the Beta otherwise password and crypt
authentication just won't work.
It puts back the loop in libpq and also fixes a couple of problems with
maintaining compatability with pre-6.3 drivers.
I haven't had final confirmation from Peter yet, but the attached patch
needs to be applied for the Beta otherwise password and crypt
authentication just won't work.
It puts back the loop in libpq and also fixes a couple of problems with
maintaining compatability with pre-6.3 drivers.
Attached is the patch to fix the warning messages from my code. I also
fixed one which wasn't my code. Apart from the usual warnings about the
bison/yacc generated code I only have one other warning message. This
is in gramm.y around line 2234. I wasn't sure of the fix.
I've also replaced all the calls to free() in gramm.y to calls to
pfree(). Without these I was getting backend crashes with GRANT. This
might already have been fixed.
This has a problem when using any authentication other than trust or
ident.
Anything using libpq will hang, because the client will go into a loop
while connecting. The following patch simply comments out two lines (a do
and a while), removing the loop. Going through the new scheme, I can't see
why this do..while loop is in there.
I've completed the patch to fix the protocol and authentication issues I
was discussing a couple of weeks ago. The particular changes are:
- the protocol has a version number
- network byte order is used throughout
- the pg_hba.conf file is used to specify what method is used to
authenticate a frontend (either password, ident, trust, reject, krb4
or krb5)
- support for multiplexed backends is removed
- appropriate changes to man pages
- the -a switch to many programs to specify an authentication service
no longer has any effect
- the libpq.so version number has changed to 1.1
The new backend still supports the old protocol so old interfaces won't
break.
I have always been under the impression that NULL is not equal to
NULL and that NULL is not equal to anything else either. If this
is the case, then this patch is correct.
If NULL _is_ equal to NULL, then I think there are other problems
in the Group By logic.
Hi -- a couple of small items concerning the January 23rd snapshot:
the inclusion of the Kerberos stuff in one Makefile, a "leading tab"
cleanup in another, and a fix for a typo in the configure script.
lock before older waiters, and having readlock people not share
locks if a writer is waiting for a lock, and waiting writers not
getting priority over waiting readers.
This is a patch to fix crashes in psql when executing queries from
an external file. The code also adds error checking to verify that
memory for "query" was allocated. The conditional for the block of
code was changed from "query == NULL" to "query_alloced == false".
The conditional, "query == NULL", was never true. This prevented
the memory being allocated for "query". A few lines later, an attempt
to write to an un-allocated memory area generated a SIGSEGV causing
the frontend to crash.
The attached patches will allow postgreSQL to compile successfully on SCO
UNIXWARE 2.1.x. The patches fix the following problems:
1. Configure did not properly recognize the UNIXWARE system as needing the
univel port. It used the sys4 port.
2. Configure did not properly process the CC flag in the template file.
3. There was no working test and set locking implementation for the native
UNIXWARE compiler.
4. The test and set locking used for Intel X86 that was selected by defining
NEED_I386_TAS_ASM could fail in a multi-processor environment.
5. The makefiles for libpq and libpgtcl did not make a shared library for
the univel port.
varchar length.
Cleans up code so attlen is always length.
Removed varchar() hack added earlier.
Will fix bug in selecting varchar() fields, and varchar() can be
variable length.
nodeAgg.c: WARN -> NOTICE for elog
parse_oper.c: was created after patch for fmgr_info, so function call wrong
scan.c: regenerated for i386_solaris using flex 2.5.4
gethostname.c: required prototype for gethostname() function
config.h.in: create prototype for isinfo() function
isinf.c: "fake" isinf() under i386_solaris using fpclass() call...
Patch by: wieck@sapserv.debis.de (Jan Wieck)
One of the design rules of PostgreSQL is extensibility. And
to follow this rule means (at least for me) that there should
not only be a builtin PL. Instead I would prefer a defined
interface for PL implemetations.
Pass List* of Aggregs into executor, and create needed array there.
No longer need to double-processs Aggregs with second copy in Query.
Fix crash when doing:
select sum(x+1) from test where 1 > 0;
OK, here comes a patch, DBD::Pg (and possibly other 3rd party clients)
can connect to unix sockets.
Patch is against current source tree.
Background:
libpq set some policy for client, which it should not
IMHO. It prevent some 3rd party clients to connect with
unix domain sockets etc.
reference to the name of the shared library, instead of dereferencing
the definition from the top of the file.
From: Tom I Helbekkmo <tih@Hamartun.Priv.NO>
==========================================
What follows is a set of diffs that cleans up the usage of BLCKSZ.
As a side effect, the person compiling the code can change the
value of BLCKSZ _at_their_own_risk_. By that, I mean that I've
tried it here at 4096 and 16384 with no ill-effects. A value
of 4096 _shouldn't_ affect much as far as the kernel/file system
goes, but making it bigger than 8192 can have severe consequences
if you don't know what you're doing. 16394 worked for me, _BUT_
when I went to 32768 and did an initdb, the SCSI driver broke and
the partition that I was running under went to hell in a hand
basket. Had to reboot and do a good bit of fsck'ing to fix things up.
The patch can be safely applied though. Just leave BLCKSZ = 8192
and everything is as before. It basically only cleans up all of the
references to BLCKSZ in the code.
If this patch is applied, a comment in the config.h file though above
the BLCKSZ define with warning about monkeying around with it would
be a good idea.
Darren darrenk@insightdist.com
(Also cleans up some of the #includes in files referencing BLCKSZ.)
==========================================
> then you try get substr, which consists only of last char in string
> you get all string
>
> For example:
> userbase=> select substr('123456', 6,1) ;
> substr
> ------
> 123456
> (1 row)
>
From Edmund Mergl <E.Mergl@bawue.de>
Comment-out dynamic link function declarations since they are all
provided by the system.
Should we bother continuing to support non-elf Linux systems??
o A new patch that contains the following changes:
-- The pg_pwd file is now cached in the postmaster's memory.
-- pg_pwd is reloaded when the postmaster detects a flag file creat()'ed
by a backend.
-- qsort() is used to sort loaded password entries, and bsearch() is
is used to find entries in the pg_pwd cache.
-- backends now copy the pg_user relation to pg_pwd.pid, and then
rename the temp file to be pg_pwd.
-- The delimiter for pg_pwd has been changed to a tab character.
and from SELECT ... INTO ... support code.
Allow NOT, IS NULL, IS NOT NULL in constraints.
Define unionall boolean flag in SubSelect structure.
Implement row descriptors: (a, b, c) = (x, y, z).
Change IS TRUE, IS FALSE, etc. to expressions using "=" rather than
function calls to istrue() or isfalse() to allow optimization.
Force type for TRUE and FALSE to bool.
Makefile.global.
End result, if all goes well, should allow for much easier porting, since
there will no longer be a concept of a "port". Most, if not everything,
*should* be determined by configure, or by the compiler itself. Still
work to be done though :)
tree "non-PORTNAME" dependent. Technically, anything that is PORTNAME
dependent should be able to be derived at compile time, through configure
or through gcc
async.c: #include <port-protos.h> surrounded by an #ifdef HAVE_STRDUP
vacuum.c: #include <port-protos.h> commented out...can someone comment as
to why it was included, as it doesn't seem to have any effect
under FreeBSD so far...would like some sort of #ifdef wrapper
like async.c if possible
Essentially, this cleans things up so that if PORTNAME isn't defined (I'm
working on getting rid of it for FreeBSD, at least, to see if its possible)
none of the PORTNAME related stuff gets passed around.
Had a little bit of -I related redundancy as well
Formerly allowed only single arguments.
Declare column constraints using the usual list mechanism rather
than explicit itemized lists.
Remove NOTNULL from default clause syntax (retain "NOT NULL").
NOTNULL is not SQL92; eventually remove it from expressions too?
Move ISNULL, NOTNULL to Postgres-specific token declarations.
Fix up tabs and indenting on new CREATE USER commands.
src. It is in the function ParseACL. When I find that I
can not allocate enough memory for the ACL structure I
return an NULL instead of doing an exit_nicely(g_conn);
From: Matthew C Aycock <maycock@scuba.pcpipeline.com>
Clean up formatting of code
Integrate new functions into dumpTable
This is not tested yet...have to recompile server due to patches from
Todd...but this compiles cleanly as it stands now
to NOTICE messages so that execution proceeds rather than halting.
These clauses are ignored as stated in the messages.
Allow NOT NULL UNIQUE syntax (both were allowed individually before).
Allow Postgres-style casting ("::") of non-constants.
btree support functions. Don't know why this was that way, but would
assume that these should be consistant with all other types with
hash support. Regression tests OK.
Change a few comments and field alignment to make things more readable.
certainly OK for datetime since it is a float8 and should be OK for
timespan since the first field within timespan is a float8.
Use float8 hash function for time type (rather than char8).
Don't know why these few were not already this way, and don't know if
there is some hidden problem with this, but assume it was done
accidentally as entries were copied from other operators.
Regression tests are OK, but...
Move one block of declaration source to keep OIDs in increasing order.
Did not change OID values, just moved source code.
Support SQL92 syntax for type coersion of strings (type 'typeval').
Example: "DATETIME 'now'". This works only for string constants and can
not replace the CAST and ::type syntax which behave identically in this
context.
Rename PG_DATESTYLE to PGDATESTYLE environment variable.
Move environment variable code to a different place so it now works!
Note that regression tests can now run with "setenv PGTZ PST8PDT"
at the frontend rather than requiring the backend to have TZ set.
My analysis of the formerly mentioned IPC reinitialization problem was
hampered by an imprecise error message. I have rewritten it so it is
clearer and more accurate.
Implement SET keyword = DEFAULT and SET TIME ZONE DEFAULT.
Re-enable JOIN= option in CREATE OPERATOR statement (damaged for v6.2).
Allow more SQL and/or Postgres reserved words as column identifiers
or, if there are shift/reduce problems, at least as column labels.
#define StrNCpy(dst,src,len) \
(strncpy((dst),(src),(len)),(len > 0) ? *((dst)+(len)-1)='\0' : \
NULL,(void)(dst))
^^^^^^ - to avoid "value computed is not used" from gcc
in ma-a-any places (should to fix thouse places instead, but ...
time)
config.h.in:
/*
* TBL_FREE_CMD_MEMORY: free memory allocated for an user query inside
* transaction block after this query is done.
*/
#define TBL_FREE_CMD_MEMORY
- this is default now.
CURRENT_USER.
Add syntax for primary and foreign keys.
Change optional syntax in CREATE INDEX to avoid parsing conflict with
TIMESTAMP WITH TIME ZONE data type (use USING <class> rather than WITH...).
Decouple various categories of data type syntax to allow the most possible
non-ambiguous extensions to SQL92 for column names and labels. This should
make the parser a bit more understandable, or at least easier to find
where and how the data types are handled.
Support syntax for IN and EXISTS clauses with subselects.
Support SQL92 syntax for IS TRUE/IS FALSE/IS NOT TRUE/IS NOT FALSE.
Allow true/false, yes/no, 1/0. Throw elog warning if anything else.
Allow shorter strings, so "t", "tr", "tru" and "true" match "true".
Old behavior accepted anything starting with "t" as TRUE,
everything else as FALSE.
rather than just 't' and 'T'. This allows yes/no and 1/0
to be interpreted as one might expect.
Clean up function declarations to use bool as the type for arguments
and return values.
lconv is already taken as a struct identifier defined in a /usr/include file.
Only has an effect with USE_LOCALE enabled.
(Apparently unique to AIX and/or AIX compiler? thomas)
This patch fixes a few results in DatabaseMetaData, and updates the README
and TODO files (the later being a new file).
The TODO file lists the things that need to be looked into after 6.2 is
released, and describes the problem with Large Objects.
I've found a problem in the Postgresql jdbc driver.
"ReceiveInteger" shifts a received byte right instead of left.
This means that only the least significant byte is read into the int.
Reviewed by: Peter T Mount <patches@maidast.demon.co.uk>
Add SQL/92 types decimal and numeric (temporary for syntax support).
These types need more support in the backend to be really implemented,
and the parser will need to be changed at that time.
Adjust limits on precision parameters for FLOAT(p) to match IEEE-compliant
arithmetic. Perhaps these limits should be processor-specific or obtained
from system include files instead.
Use qsort to sort array of tuples for nextrun when current
run is done and put into leftist tree from sorted array!
It's much faster and creates non-bushy tree - this is ve-e-ery good
for perfomance!
2. Limit number of tuples in leftist trees:
- put one tuple from current tree to disk if limit reached;
- end run creation if limit reached by nextrun.
3. Avoid mergeruns() if first run is single one!
Add parsing for UNION and outer JOINs.
Implement SQL92 "WITH TIME ZONE".
Allow some reserved words as identifiers and column labels.
Clean up indentation and "orphan spaces and tabs".
Implement extended comments ("/* ... */") using exclusive states.
Modify definitions of operators to remove some restrictions on characters
and character order.
Change box terminology from "length" to "width".
Use length terminology in common with other geometric types (usually perimeter).
Fix bugs in line arithmetic which resulted in bad intersection calculations.
Deprecate temporary unstored slope fields.
Check explicitly for intersections at endpoints to avoid rounding ugliness.
Add center() routines for lseg, path, polygon.
Add distance() routines for circle-polygon, polygon-polygon.
Check explicitly for points and polygons contained within polygons
using an axis-crossing algorithm. (Old code just checked bounding boxes).
Add routine to convert circle-box.
*whew*
This matches the behavior of the original formatting for abstime.
Repair datetime + timespan date arithmetic for year boundaries.
From patch submitted by Dave Skinner.
the problem only manifests itself when adding years/months and hours
when the hours:minutes:seconds pushes over midnight.
Fix interpretation of times with explicit timezone when the timezone is
in daylight savings time and is not the default timezone.
Allow interpretation of explicit timezone when it is specified as two words:
<standard time> DST". For example, "MET DST" (Middle European Time Daylight
Savings Time). This syntax is found in the zic package on Linux boxes at least.
Subject: [PORTS] Patches for Irix 6.4
I have worked out how to compile PostgreSQL on Irix 6.4 using the -n32 compiler
mode and version 7.1 of the C compiler. (The n32 compiler use 32 bits
addressing,
but allows access to all the instructions in the MIPS4 instruction set.)
There were several problems:
1) The ld command is not referenced as a macro in all the Makefiles. On
this platform, you have to include -n32 on all the ld commands. Makefiles
were changed as needed.
3) Lots of warnings are generated from the compiler. Since the regression
tests worked OK, I didn't attempt to fix them. If anyone wants the compilation
log, please let me know, and I'll email it to you.
The version of postgresql was 970602. Here is Makefile.custom:
CUSTOM_COPT = -O2 -n32
MK_NO_LORDER = 1
LD = ld -n32
CC += -n32
Subject: [PATCHES] sequences display in psql
Well, I am away at Progress training (not Postgres!!) and desided to do
this patch during a break. This will allow listing of sequences in
addition to listing of tables and indicies:
\d would should indicies, tables, and sequences
\ds would show sequences only.
Subject: [PATCHES] DG/UX 5.4R11 patches
1) config.guess -- it doesn't understand that the new default
"TARGET_BINARY_INTERFACE" (m88kdgux) is now an ELF
format, not BCS.
Subject: [PATCHES] psql and large objects fix
Psql was broken by using "Inv[0-9]+" instead of "xin[xv][0-9]+" to not
show large object files. Been this way for a long time too. Relic of
an older naming convention, perhaps?
Subject: [PATCHES] backend/lib/fstack.c
- The FixedStackIsValid so obviously needs to be a macro
- FixedStackContains only called if assert checking on
Subject: [PATCHES] More psql and libpq patches
Well..these would be the last patches until the release (I hope)...
I ran the regression tests while watching psql under purify, and it did
not leak even one byte.
In this patch:
* Plugged a major leak when PSQL reads files for input (either through
\i options or through -f option)
* Fixed the one remaining leak in PSQL in not clearing PGresult *results
everywhere it is supposed to. (Thanks Tymm)
* Fixed A small leak in PSQL not clearing all the PGsettings correctly.
* A not-so-obvious (but small) leak in Libpq when PQsetdb fails for any
reason.
* Added \n to some Libpq error messages to make them easier to digest..
* Finally, added /* PURIFY */ comment to some of the code indicating
the reason for why it was added/changed...for future developers.
Subject: [HACKERS] src.original/./backend/lib/fstack.c
Another change I suggested. I bracket an unused function and add a
return to quiet the compiler. In addition I added an internal
consistency check.
Subject: [HACKERS] backend/optimizer/geqo/geqo_erx.c
I sent these changes in with a bunch of others. Some were folded in but
others, like these, were not. I am not sure why so I am resending this
to the developers list by itself for discussion.
The readon why I suggest these changes is that the compiler can't tell
that minimum_count is initialized before it is used. The tests that I
add in here will cause an immediate error if it doesn't. As the comments
below suggest, if it is 100% guaranteed that the variable will always
be initialized then how this is so should be commented here. I don't
know how much strain the actual test puts on the performance but if it
isn't too much then maybe leave it in for absolute safety anyway.
There are also a few returns just to stop warnings.
for join-relations. Sizes already computed by
prune_rel_paths():compute_joinrel_size().
joinrels.c:
< if ( _use_right_sided_plans_ )
---
> if ( _use_right_sided_plans_ &&
> length (outer_rel->relids) > 1 )
- r_plans are useful when outer_rel is join-relation... It
decreases the size of search space...
psql.c: In function `HandleSlashCmds':
psql.c:1141: warning: `optarg3' might be used uninitialized in this function
psql.c:1157: warning: `optarg3' might be used uninitialized in this function
-> char *optarg3 = NULL;
start time equal to tuple->t_tmax.
Privent shrinking if there are tuples modifyed by running transactions
(it concerns system relations only, currently).
Subject: [PATCHES] patch for a memory leak
Well...I screwed up and posted the wrong patch for psql originally..
The patch for that patch wposted below will fix it..
Subject: [PATCHES] Another destroydb patch
This is a patch to my previous destroydb patch cause some people wanted
slightly different behavior. After this patch is applied, destroydb
will destroy a database as usual, but if added -i flag (which could be
aliased like rm -i) would ask for confirmation.
Subject: [PATCHES] pg_dump memory leak patch
This patch fixes a HUGE memory leak problem in pg_dump.
Pretty much anything that was allocated was never freed and Purify
reported about 40% possible memory leak and 6% actual leak. I added
functions to clear out all the allocated structures. After the patch
Purify returns 0 for number of bytes leaked...
Subject: [PATCHES] psql - \dt,\di commands.
I sent this a couple of months ago in re a request by Maxim
Kozin, but I had the patch reversed, creating some confusion
over applying it.
Here's a more complete version.
Adds \dt to list only tables/views and \di to list only
indicies. \d will still work as before.
Subject: [PATCHES] destroydb patch
I am including a patch for destroydb to ask for confirmation before
deleting databases (after I accidentally deleted mine)...destroydb -y
would force delete without any confirmation.
Subject: [PATCHES] memory leak patches in libpq and psql
A couple of small memory leak patches (detected with Purify) primarily
in libpq.
* Fixed (NULL) border problem in psql (run psql, do \m, then select
something from a table...row separators will be nulls)
* Fixed memory leak with the abovementioned border not being freed
properly.
* Fixed memory leak in freePGconn() not freeing conn->port
* Fixed up PQclear() to free parts of PGresult only if these
parts are not null.
* Fixed a decent memory leak that occured after executing every command
in psql. PGresult *results was not freed most of the time.
There is still a leak being detected (2 bytes) in readline functions, but
I think this is old readline library. I will install new one and test it.
Subject: [PATCHES] pqcomprim.c patch
This is the patch by Robert Bruccoleri to fix the endian problem.
(Actually, it's the reverse of his patch. He must have gotten the
order wrong.)
/*
* RelationFlushRelation () below will flush relation information
* from the cache. We must call smgrclose to flush relation
* information from SMGR & FMGR, too. We assume that for temp
* relations smgrunlink is already called by heap_destroyr
* and we skip smgrclose for them. - vadim 05/22/97
*/
smgrclose(reln->rd_rel->relsmgr, reln);
- it avoids memory leaks in SMGR & VFD.
RelationFlushRelation():
there is no more call FileInvalidate(RelationGetSystemPort(relation));
- invalid (FileInvalidate() expects File, not SMGR' fd)
- unuseful anyway.
mdunlink() and mdclose() (too !!!) now free MdfdVec for relation
and add it to free list, so it may be re-used for another relation
later.
2. Fix VFD-manager memory leak (found by Massimo ... and me):
mdunlink() has to call FileUnlink() to free allocation for fileName
and add the Vfd slot to the free list.
/*
** You can have as many strategies as you please in GiSTs, as
** long as your consistent method can handle them
*/
#define GISTNStrategies 100
^^^
- too big number:
strat.h->StrategyEvaluationData->StrategyExpression expression[12]
^^
- so 12 is real max # of strategies, or StrategyEvaluationIsValid
crashes backend (called if CASSER defined).
To: pgsql-patches@postgreSQL.org
Subject: [PATCHES] DROP AGGREGATE gram.y typo...
Somehow I dropped a comma in the gram.y part (line 129) of my
patch for drop aggregate. Here's a correct patch for gram.y.
PS. I hope I got the right comma, manually applied :) (scrappy)
Subject: [PATCHES] AIX make patch resubmitted.
Misc patches for AIX from Darren:
1) New src/makefiles/Makefile.aix This patch should only be
applied if the following patch (4) is applied to backend/Makefile!
Still looking into having configure determine the last line to do
the shared link. The 325 code will work for 41, so I put that in
as the default. Included a commented out 41 line for completeness.
*and*
4) Patch the backend Makefile. I've reviewed this patch with respect to the
other ports that use MAKE_EXPORTS (svr4 and univel) as closely as I could
and I don't see where it will break them. If it does, please let me know
and I'll rework it somehow.
Subject: [PATCHES] Re: [PORTS] AIX 6.1 fixes...
Here are the patches for the two things that wouldn't make it thru the AIX
compiler. The geo_ops.c change is harmless I believe. The nbtcompare.c patch
fixes me, but I don't know about any other ports. Maybe wait on that one
until Vadim decides what to do about the unsigned vs signed chars varlena
issue.
all local buffers @ xact commit, so accordingly nextFreeLocalBuf
is first local buffer now.
It helps to avoid unnecessary local buffer allocations in LocalBufferAlloc()
latter ("memmory leaks" in 'order by').
2. ResetLocalBufferPool() lost allocated local buffers:
memset(LocalBufferDescriptors, 0, sizeof(BufferDesc) * NLocBuffer);
(local buffers leak @ xact aborts).
Bring optional new-storage date and time up to date and test.
This new storage format should fix the "Sparc gcc -O2 bug".
(Enable new code with USE_NEW_DATE and USE_NEW_TIME in dt.h)
Subject: [PORTS] minor fix for DGUX port
src/include/port/dgux.h needs the following three lines appended:
#ifndef BYTE_ORDER
#define BYTE_ORDER BIG_ENDIAN
#endif
I believe this to be correct for DG/UX on M88k processors. I don't have one of
the new Intel-based boxes to check on.
Include some additional path functions which were coded but omitted here.
Add translation and rotation/scaling operators for some geometric types.
Fix bugs in some geometry comparison operator declarations.
Add type conversion functions for floating point numbers.
Check for zero in unary minus floating point code (IEEE allows an
explicit negative zero which looks ugly in a query result!).
Ensure circle type has non-negative radius.
Subject: [PATCHES] libpq patch
Hi,
here is a small patch which fixes two problems:
1. libpq/libpq-fe.h:
somehow disappeared the line
#define DefaultOption ""
now compilation stops with an error complainig an
unknown DefaultOption (970508).
2. Same patch as I sent already twice, but it never made it
into the source tree: there is no default value for
AuthType and Password. This way any libpq-application
(i.e. perl-scripts) which use the function PQconnectdb
will break with PostgreSQL-6.1. The patch simply uses
an empty string as default value.
Subject: [PATCHES] Patches for boolean, timespan and reltime regression tests.
Hi All,
Here are a couple of patches to the regression tests to introduce
some specific ordering to the results.
I've only made changes to the queries that were exhibiting differences
on my regression runs.
This will also have the side effect of testing the ordering code for
the boolean and some of the time types.
Subject: [PATCHES] libpq SET var TO patch
One last, I hope. This one corrects a bogus format string, and
actually sends the contents of PG_DATESTYLE to the backend. That
means, you can do a setenv PG_DATESTYLE 'iso', and your libpq
will pick that up and tell the backend.
Subject: [PATCHES] port patch: ultrix4
ultrix4 doesn't compile without this. this also fixes a problem
with dynamic loading (ultrix relocatable objects must be loaded
with -G 0).
Subject: [PATCHES] Three small patches.
Hi,
Here are 3 small patches to the postgreSQL source sup'd on
the 6th May 1997.
The 1st 2 fix the shell backslash "c" handling used to suppress
the newline on some unix shells. (The \c needs to be inside quote.)
The 3rd may or may not be the correct way to fix the missing
define of INDEX_MAX_KEYS in pg_dump.h
fd = FileNameOpenFile(path, O_RDWR|O_CREAT|O_EXCL, 0600);
/*
* If the file already exists and is empty, we pretend that the
* create succeeded. During bootstrap processing, we skip that check,
* because pg_time, pg_variable, and pg_log get created before their
* .bki file entries are processed.
*
> * As the result of this pretence it was possible to have in
> * pg_class > 1 records with the same relname. Actually, it
> * should be fixed in upper levels, too, but... - vadim 05/06/97
> */
when btree used in innerscan with run-time key which value
passed by pointer.
Fix: keys ordering stuff moved to _bt_first().
Pointed by Thomas Lockhart.
Subject: [HACKERS] Inputting money
I notice that I have to put single quotes around money amounts if there
is a decimal point in the value. I appears to be happening because there
is something changing things like "123.45" to "123.450000" and the code
has a problem with that. There may be a better way to fix this but here
is a simple change to cash.c that lets it accept trailing zeroes.
Add mixed-case #define synonyms to avoid changing more source code.
Add comparison operators for boolean.
Add aggregate min() and max() for datetime and timespan.
Add comparison operators to boolean and smaller/larger operators to datetime
and timespan. Fix int4 overflow math problem in timespan comparison operators.
Subject: [PATCHES] to make regress.sh shell friendly to echo.
Hi,
I needed to make the following change to regress.sh to make it more
shell friendly.
The Solaris /bin/sh, and others, use \c to supress the newline.
the DROP TABLE calls from the destroy.sql file to the 'types' .sql files,
so that they are self-contained
btree_index, hash_index and misc all fail as there seems to be missing
a 'misc.out' expected file...have asked Thomas for one...
=============== destroying old regression database... =================
=============== creating new regression database... =================
=============== running regression queries... =================
create_function_1 .. ok
create_type .. ok
create_table .. ok
create_function_2 .. ok
Here are patches which should help fix timezone problems in the
datetime and abstime code. Also, I repatched varlena.c to add in
some comments and a little error checking on top of Vadim's earlier
repairs. There are slight mods to the circle data type to have the
distance operator between circles measure the distance between
closest points rather than between centers.
Subject: [PATCHES] Patches for compiling 6.1 on Digital Unix 3.2c
Attached to this message are the patches I needed to compile 6.1 cleanly
under Digital Unix 3.2c with DEC cc.
I hope these are the last ones. At least, the number of files needing a
patch has decreased noticeably since I sent my previous patches. Nice work
:-)
One of the patches is a bug fix, but I'm including it here anyway.
With these patches applied, the beast seems to work properly. However,
I've done only some preliminary tests. More on this later (but hopefully
before the April 30 deadline... :-)
postgres backend processes end up as so called zombies. It seems that
only Linux a.out (libc.4.6.27) systems are affected.
By:
Wolfgang Roth <roth@statistik.uni-mannheim.de>
nestloop's join clauses doesn't work in some cases:
* 1. fix_indxqual_references may change varattno-s in
* inner_indxqual;
* 2. clauses may be commuted
Subject: Re: [PATCHES] SET DateStyle patches
On Tue, 22 Apr 1997, Thomas Lockhart wrote:
> Some more patches! These (try to) finish implementing SET variable TO value
> for "DateStyle" (changed the name from simply "date" to be more descriptive).
> This is based on code from Martin and Bruce (?), which was easy to modify.
> The syntax is
>
> SET DateStyle TO 'iso'
> SET DateStyle TO 'postgres'
> SET DateStyle TO 'sql'
> SET DateStyle TO 'european'
> SET DateStyle TO 'noneuropean'
> SET DateStyle TO 'us' (same as "noneuropean")
> SET DateStyle TO 'default' (current same as "postgres,us")
>
> ("european" is just compared for the first 4 characters, and "noneuropean"
> is compared for the first 7 to allow less typing).
>
> Multiple arguments are allowed, so SET datestyle TO 'sql,euro' is valid.
>
> My mods also try to implement "SHOW variable" and "RESET variable", but
> that part just core dumps at the moment. I would guess that my errors
> are obvious to someone who knows what they are doing with the parser stuff,
> so if someone (Bruce and/or Martin??) could have it do the right thing
> we will have a more complete set of what we need.
>
> Also, I would like to have a floating point precision global variable to
> implement "SET precision TO 10" and perhaps "SET precision TO 10,2" for
> float8 and float4, but I don't know how to do that for integer types rather
> than strings. If someone is fixing the SHOW and RESET code, perhaps they can
> add some hooks for me to do the floats while they are at it.
>
> I've left some remnants of variable structures in the source code which
> I did not use in the interests of getting something working for v6.1.
> We'll have time to clean things up for the next release...
Subject: [PORTS] Configure for DEC-Alpha
Configure script properly detects alpha-dec-osf4.0 machine, but
sets a default GENERIC template for it. I modified tempplate/.similar to
add alpha-dec-osf4.0=alpha. Then configure properly set the template to
alpha.
Subject: [PATCHES] Patch for configure.in to not ask for CASSERT
The following patch defaults to CASSERT, so it doesn't ask you. You can
still use --enable-cassert and --disable-cassert to do it explicitly.
Default: disabled
Subject: [PATCHES] date/time timezone patches (mail bounced?)
Here are some hacks to get timezone behavior for the various time
data types to be compatible with v6.0. Although we have some hooks
already installed to get timezone info from the client to the
server, it still isn't clear if that can correctly transfer enough
timezone info to make the behavior the same as if timezone info
were derived from the server as is now the case. We certainly
won't resolve it in a day, so I think we are stuck with server-only
timezones for v6.1.
OK, here are a passel of patches for the geometric data types.
These add a "circle" data type, new operators and functions
for the existing data types, and change the default formats
for some of the existing types to make them consistant with
each other. Current formatting conventions (e.g. compatible
with v6.0 to allow dump/reload) are supported, but the new
conventions should be an improvement and we can eventually
drop the old conventions entirely.
For example, there are two kinds of paths (connected line segments),
open and closed, and the old format was
'(1,2,1,2,3,4)' for a closed path with two points (1,2) and (3,4)
'(0,2,1,2,3,4)' for an open path with two points (1,2) and (3,4)
Pretty arcane, huh? The new format for paths is
'((1,2),(3,4))' for a closed path with two points (1,2) and (3,4)
'[(1,2),(3,4)]' for an open path with two points (1,2) and (3,4)
For polygons, the old convention is
'(0,4,2,0,4,3)' for a triangle with points at (0,0),(4,4), and (2,3)
and the new convention is
'((0,0),(4,4),(2,3))' for a triangle with points at (0,0),(4,4), and (2,3)
Other data types which are also represented as lists of points
(e.g. boxes, line segments, and polygons) have similar representations
(they surround each point with parens).
For v6.1, any format which can be interpreted as the old style format
is decoded as such; we can remove that backwards compatibility but ugly
convention for v7.0. This will allow dump/reloads from v6.0.
These include some updates to the regression test files to change the test
for creating a data type from "circle" to "widget" to keep the test from
trashing the new builtin circle type.
Subject: [HACKERS] Another patch to configure.in
I heard very little in objections/approvals to defaulting some of the
parameters to configure. Enclosed is a patch to configure.in which
removes the questions for
PGPORT
USE_LOCALE
NOHBA
By default (i.e. assuming you don't put anything extra in the configure
command line), it assumes PGPORT=5432, USE_LOCAL=no and NOHBA=no (i.e.
HBA is turned on)
--with-pgport=PGPORT_NO Over-rides the PGPORT value
--enable-locale enables USE_LOCALE
--disable-hba disables HBA
Just for completeness:
--prefix=BASEDIR Defaults to /usr/local/pgsql
--with-template=TEMPLATE Defaults to asking you
Subject: [PATCHES] 970417: some large object patches
Two patches here, made against 970417. Both have to do with large
objects:
1. lobjfuncs was not initialized in PQconnectdb. This causes
failure later if large objects are used. (Someone already
caught this error in PQsetdb.)
2. Postgres functions lo_import and lo_export sometimes
produce garbage for the file names because the filename
strings aren't always terminated by \0. (VARDATA isn't
necessarily null terminated.)
Subject: [PATCHES] 970417: two more patches for large objects
Here are two more patches:
1. pg_getint doesn't properly set the status flag when
calling pqGetShort or pqGetLong. This is required when
accessing large objects via libpq. This, combined with
problem 1 above causes postgres to crash when postgres
tries to print out the message that the status was not
good.
2. ExceptionalCondition crashes when called with detail =
NULL. This patch prevents dereferencing the NULL.
Subject: [HACKERS] Patch: set date to euro/us postgres/iso/sql
Here a patch that implements a SET date for use by the datetime
stuff. The syntax is
SET date TO 'val[,val,...]'
where val is us (us dates), euro (european dates), postgres,
iso or sql.
Thomas is working on the integration in his datetime module.
I just needed to get the patch out before it went stale :)
table. The table name is de-allocated by the CommitTransactionCommand()
in vc_init() before it is copied in VacRel.data and sometimes this causes
a SIGSEGV. My patch simply moves the strcpy before vc_init.
Submitted by Massimo Dal Zotto <dz@cs.unitn.it>.
index tuple (logical position within A LEVEL). bti_oid & bti_dummy
taken off from BTItemData.
2. Fix for multi-column indices (nbtsearch.c):
_bt_binsrch() - for searches on internal pages having keysize <
number of attrs we point at the last item < the scankey, not at the
first item = the scankey;
_bt_moveright() - if keysize < number of attrs we compare scankey with
_last_ item on current page to decide should we move right or
not.
Subject: [HACKERS] Money integration patches
Here are patches to integrate the money data type. I have included
some math and aggregate functions and have made the locale support optional
by #ifdef USE_LOCALE bracketing of functions.
Modules affected are:
builtins.h.patch
cash.c.patch
cash.h.patch
main.c.patch
pg_aggregate.h.patch
pg_operator.h.patch
pg_proc.h.patch
pg_type.h.patch
I changed the data type to be pass-by-reference rather than by-value
to pave the way for a larger internal representation (64-bit ints?).
Also, I changed the tabbing of cash.c and cash.h to match most of
the other Postgres source code files (4 space indent, 8 spaces == 1 tab).
The locale stuff should be tested under another convention (Russian?)
but I don't know what the correct results should be so perhaps someone
else can give them a try. Will update docs and regression tests in
the next few days.
invalid macro definitions, the compiler complains about:
"pqcomprim.c", line 48.9: 1506-275 (S) Unexpected text ';' ignored.
"pqcomprim.c", line 61.9: 1506-275 (S) Unexpected text ';' ignored.
The ';' terminating the macro definition ntoh_s(n) on line 27 and
ntoh_l(n) on line 28 should be removed.
Pointed out by: Olaf Mittelstaedt <MSTAEDT@va-sigi.va.fh-ulm.de>
Makefile.global and move them to seperate 'include' makefiles
Over time, should become even more port specific:
ie. Makefile.BSD44_derived should be broken down into netbsd/freebsd
specific ports
pg_proc.h still needs modifying, but this gets it in there so that we can
get around any compiler bugs. Will try and get the pg_proc.h entries done
up later tonight...
Subject: Re: [HACKERS] GEQO and views (rules)
Oke, this was caused by a classic bug :-/
I thougth, root->base_relation_list_ could be
represented as relid string 1-2-3-4- etc.
Instead, in case of views, the count of relids doesn't start with "1" but
maybe 4-5-6- etc . :-(
GEQO patch follows ... views are now all right.
2. PageWeights are variables now.
3. Fixed using ceil((double)selec*indextuples) as estimation
for expected heap pages: ceil((double)selec*relpages) now.
use sum(npages)/((nkeys == 1) ? 1 : nkeys + 1) as expected index page
estimation for multi-key quals - instead of sum(npages).
In old code npages for x > 10 and x < 20 is twice as for x > 10 - cool ?
'master' file
Commit mods to regress.sh so that split out tests are run...look forward
to finding out how to do a proper redirect to continue visual cleanup :)
Subject: [HACKERS] Fix for European dates
This apparently fixes the European date reading problem reported
by several (European) bleeding edge adopters. I tried a few test
cases and it doesn't break the non-EuroDate cases in my test suite.
FreeBSD
The Makefile(s) have all been cleaned up such that there is a single
LDFLAGS vs LD_ADD or LDADD or LDFLAGS or LDFLAGS_BE. The Makefile(s)
should be alot more straightforward then they were before...and
consistent
Further extended Makefile.global/build/configure so that we can
have a 'template' file for each OS (and each version of OS, as in BSDi)
which is used as much as possible to generate Makefile.global
Any future ports should look at using the template file as a basis,
before moving over to Makefile.global.
This will most probably break alot of the ports, atho I've tried to
be very neat about it...
Remove USE_LOCALE from Makefile.global.in
Add USE_LOCALE to build/configure/config.h
Add check for BUILDRUN in configure to make sure that build is run before
configure
Subject: [HACKERS] timestamp.c changes
I sent in changes previously and they were rejected because they didn't
follow ANSI spec. Here is the input part of the changes again. Even
though it allows more flexibility for inputting different formats, it
is also backwards compatible with the standard version. I have also
not changed the output format so it will still output the ANSI forms.
Is this acceptable to everyone?
Subject: [HACKERS] Aggregate function patches
Here are the aggregate function patches I originally sent in last December.
They fix sum() and avg() behavior for ints and floats when NULL values are
involved.
I was waiting to resubmit these until I had a chance to write a v6.0->v6.1
database upgrade script to ensure that existing v6.0 databases which have
not been reloaded for v6.1 do no break with the new aggregate behavior.
These scripts are included below. It's OK with me if someone wants to do
something different with the upgrade strategy, but something like this
was discussed a few weeks ago.
Also, there were a couple of small items which cropped up in doing a clean
install of 970403 (actually 970402 + 970403 changes since the full 970403
tar file appears to be damaged or at least suspect). They are the first
two patches below and can be omitted if desired (although I think they
aren't dangerous :).
Subject: [HACKERS] More date time functions
Here are some additional patches mostly related to the date and time
data types. It includes some type conversion routines to move between
the different date types and some other date manipulation routines such
as date_part(units,datetime).
I noticed Edmund Mergl et al's neat trick for getting function overloading
for builtin functions, so started to use that for the date and time stuff.
Later, if someone figures out how to get function overloading directly
for internal C code, then we can move to that technique.
These patches include documentation updates (don't faint!) for the built-in
man page. Doesn't yet include mention of timestamp, since I don't know
much about it and since it may change a bit to become a _real_ ANSI timestamp
which would include parser support for the declaration syntax (what do you
think, Dan?).
The patches were developed on the 970330 release, but have been rebuilt
off of the 970402 release. The first patch below is to get libpq to compile,
on my Linux box, but is not related to the rest of the patches and you can
choose not to apply that one at this time. Thanks in advance, scrappy!
Subject: [HACKERS] Patch: SET var TO 'val'
Here is a patch that adds a "SET variable TO 'somevalue'" capability
to the parser, and then calls the SetPGVariable() function (which does
just issue a elog(NOTICE) to see whether it works).
That's the framework for adding timezone/date format/language/...
stuff.
Subject: [HACKERS] Small patch to pgtclCmds.c
Hi I have made the following small change to the extensions I made to
pgtclCmds.c quite a while ago.
At the moment there is a -assignbyidx option to pg_result assigning the
returned tuples to an array by using the 1st field of the select statement
as the key to the array.
eg "select name,age from vitalstatistics" will result in an array with
myarray(peter) = 32
myarray(paul) = 45
Often I need to have a pseudo-multi dimentional
array eg. "select name,age from vitalstatistics where occupation='plummer'
I would like to be able to generate an array
newarray(peter,overpaid) = 32
So to add a arbitrary string to the key value I have extended
pg_result $res -assignbyidx $arrayname
to have an optional argument
pg_result $res -assignbyidx $arrayname $appendstr
So that that string is appended to the key value.
Subject: [HACKERS] locale patches !
Hi there,
here are little patches to get Postgres 6.1 works with locale stuff.
This is a patch against 970402.tar.gz, there are no problem to apply them
by hand to 6.0 release. Collate stuff tested about 1-2 months in real
working database but I'm sure there must be no problem. US hackers
could vote against locale implementation ( locale for sure will affect to
speed of postgres ), so I introduce variable USE_LOCALE which
controls locale stuff. Non-US users now could use ~* operator
for searching and <order by> for strings with nation alphabet.
Please, don't forget, as I did first time, to set environment variable
LC_CTYPE and LC_COLLATE because backend get locale information from them.
I start postmaster from a little script, assuming that shell is Bash shell
it looks like:
#!/bin/sh
export LC_CTYPE=koi8-r
export LC_COLLATE=koi8-r
postmaster -B 1024 -S -D/usr/local/pgsql/data/ -o '-Fe'
Subject: [HACKERS] Small date patches (resubmitted)
Here a some small patches for the date/time code. They set the default
output format for the datetime type to the traditional Postgres
style, and fix a date debugging declaration. I submitted these
a couple of days ago, but they might have gotten lost...
NOTE: the second patch to dt.c is what I believe D'Arcy submitted as well,
that I claimed was taken out...sorry D'Arcy, my fault :(
Subject: Re: [HACKERS] abstime "now" broken
Yes, I broke 'now' :( with an attempt at a bug fix involving
servers running in the UTC/GMT timezone. These patches fix
the problem, and have been tested in GMT (+00 hours),
PST (-08), and NZT (+12) timezones which exercized the code for
various cases including across day boundaries. btw, this code
fixes the same type of problem for 'today', 'yesterday', 'tomorrow',
for DATETIME, ABSTIME, DATE and TIME types.
The bugfix itself is quite small, but I have accumulated other
changes in the datetime data type and include them here also.
One set of changes involves printing ISO-formatted dates and
is in response to the helpful information from Kurt Lidl regarding
ANSI SQL dates. I'll send another e-mail sometime soon discussing
more issues he has raised...
Reply-To: hackers@hub.org, Dan McGuirk <mcguirk@indirect.com>
To: hackers@hub.org
Subject: [HACKERS] tmin writeback optimization
I was doing some profiling of the backend, and noticed that during a certain
benchmark I was running somewhere between 30% and 75% of the backend's CPU
time was being spent in calls to TransactionIdDidCommit() from
HeapTupleSatisfiesNow() or HeapTupleSatisfiesItself() to determine that
changed rows' transactions had in fact been committed even though the rows'
tmin values had not yet been set.
When a query looks at a given row, it needs to figure out whether the
transaction that changed the row has been committed and hence it should pay
attention to the row, or whether on the other hand the transaction is still
in progress or has been aborted and hence the row should be ignored. If
a tmin value is set, it is known definitively that the row's transaction
has been committed. However, if tmin is not set, the transaction
referred to in xmin must be looked up in pg_log, and this is what the
backend was spending a lot of time doing during my benchmark.
So, implementing a method suggested by Vadim, I created the following
patch that, the first time a query finds a committed row whose tmin value
is not set, sets it, and marks the buffer where the row is stored as
dirty. (It works for tmax, too.) This doesn't result in the boost in
real time performance I was hoping for, however it does decrease backend
CPU usage by up to two-thirds in certain situations, so it could be
rather beneficial in high-concurrency settings.
Subject: [HACKERS] backend/utils/adt/timestamp.c
Back to this timezone stuff. The struct tm has a field (tm_gmtoff) which
is the offset from UTC (GMT is archaic BTW) in seconds. Is this the
value you are looking for when you use timezone? Note that this applies
to NetBSD but it does not appear to be in either ANSI C or POSIX. This
looks like one of those things that is just going to have to be hand
coded for each platform.
Why not just store the values in UTC and use localtime instead of
gmtime when retrieving the value?
Also, you assume the time is returned as a 4 byte integer. In fact,
there is not even any requirement that time be an integral value. You
should use time_t here.
The input function seems unduly restrictive. Somewhere in the sources
there is an input function that allows words for months. Can't we do
the same here?
There is a standard function, difftime, for subtracting two times. It
deals with cases where time_t is not integral. There is, however, a
small performance hit since it returns a double and I don't believe
there is any system currently which uses anything but an integral for
time_t. Still, this is technically the correct and portable thing to do.
The returns from the various comparisons should probably be a bool.
The first fixes a warning from gcc about the assignment within the condition.
The extra set of parens should not make a difference, but with -Werror, they
are necessary.
The second fixes an "ln -s" invocation that assumes the current directory is
implicitly the target if not specified. Not true in all cases, and again, it
should not make a difference except to those implementation that it does.
From: "Michael P. Snyder" <msnyder@hawkeye.huntersmoon.com>
of endian.h. I figure that if it exists it's pretty sure that it has
the byte order information and we may catch some other ports without
any further testing.
From: "D'Arcy J.M. Cain" <darcy@druid.net>
Subject: [HACKERS] More patches for date/time
I have accumulated several patches to add functionality to the datetime
and timespan data types as well as to fix reported porting bugs on non-BSD
machines. These patches are:
dt.c.patch - add datetime_part(), fix bugs
dt.h.patch - add quarter and timezone support, add prototypes
globals.c.patch - add time and timezone variables
miscadmin.h.patch - add time and timezone variables
nabstime.c.patch - add datetime conversion routine
nabstime.h.patch - add prototypes
pg_operator.h.patch - add datetime operators, clean up formatting
pg_proc.h.patch - add datetime functions, reassign conflicting date OIDs
pg_type.h.patch - add datetime and timespan data types
The dt.c and pg_proc.h patches are fairly large; the latter mostly because I tried
to get some columns for existing entries to line up.
nicer. Also, I grabbed my copy of the Informix manual, and
added a couple of variables that make sense (formats for
money, time, a language setting, a timezone).
- New functions SetPGVariable() and GetPGVariable() in tcop/*.
These don't actually do anything for the moment, but should
be enough to implement the SET var_name TO var_val in the
parser?
SetPGVariable() expects just two strings, the var_name and
the var_value from above, and is expected to do the right thing.
Returns TRUE if everything okay.
From: "Martin J. Laubach" <mjl@wwx.vip.at>
Actually required by multi-column indices support.
We still don't use btree for 'A is (not) null', but
now btree keep items with NULL attrs using single rule
for placing/finding items on pages:
NULLs greater NOT_NULLs and NULL = NULL.
+ Bulkload code (nbtsort.c) support for multi-column indices
building and NULLs.
+ Fix for btendscan()->pfree(scanopaque) from Chris Dunlop.
Subject: [HACKERS] backend/utils/adt/nabstime.c
There is a problem with some of the calls to strftime. The second arg is
missing. In all cases the buffer is CTZName which, according to the
file init/globals.c, is char CTZName[8] so I have added this value.
I know there should be a #define set up for this but I wasn't sure
which header to put it in.
Subject: [HACKERS] libpq/pqcomm stuff and Solaris byte order
I decided to go ahead with the required changes since no one else seems
to. I don't guarantee that it is perfect but with these changes the
package actually compiles. While I was at it I added to the Sparc
Solaris header to define the byte order. Note that NetBSD sets this
in the system headers so it wasn't required there.
In particular, someone may want to check whether I removed the correct
84 lines from backend/libpq/pqcomprim.c.
Subject: [HACKERS] auth.c for kerberos.
I made pgsql with eBones(international version of Kerberos4). The
following modification was needed. And I added read permition for
group to srvtab instead of running postmaster as root.
According to man page under FreeBSD for sys_errlist[], strerror() should be
used instead...not sure if this will break other systems, so only changing
two files for now, and we'll see what "errors" it turns up
of common routines in pqcomprim.c (pq communication primitives).
Not all adapted to it yet, but it's a start.
- Rewritten some of those routines, to write/read bigger chunks of
data, precomputing stuff in buffers instead of sending out byte
by byte.
- As a consequence, I need to know the endianness of the machine.
Currently I rely on getting it from machine/endian.h, but this
may not be available everywhere? (Who the hell thought it was
a good idea to pass integers to the backend the other way around
than the normal network byte order? *argl*)
- Libpq looks in the environment for magic variables, and upon
establishing a connection to the backend, sends it queries
of the form "SET var_name TO 'var_value'". This needs a change
in the backend parser (Mr. Parser, are you there? :)
- Currently it looks for two Env-Vars, namely PG_DATEFORMAT
and PG_FLOATFORMAT. What else makes sense? PG_TIMEFORMAT?
PG_TIMEZONE?
From: "Martin J. Laubach" <mjl@wwx.vip.at>
Subject: [HACKERS] Patches for 970316 compilation
I made a small pre-emptive change in the new datetime code to eliminate
calls to infnan(). Hopefully this will make Solaris (and probably other
non-GNUlib) systems happier. Didn't find fe-connect.h in the 970316
distribution, so made one up. Also, one of the test routines needs an
update for the geo-decls.h -> geo_decls.h name change.
Patches appear below...
Subject: [HACKERS] Patch for io routines
I am currently trying to improve on the front-backend communication
routines; and noticed that lots of code are duplicated for libpq and
the backend. This is a first patch that tries to share code between
the two, more to follow.
mjl
Subject: [HACKERS] lock debug trace
This is an update to my previous patches for lock debugging, already applied
to the current sources. It adds some improvements in the output messages and
some more output in WaitOnLock(). I have used with success to trace a nasty
deadlock condition on pg_listener.
> Please apply them to the direcory "backend/optimizer/geqo".
> Two new files with different crossover techniques are included.
> Standard procedure is optimization by means of "geqo_erx.c"
> (Edge Recombination Crossover).
From: "Martin S. Utesch" <utesch@aut.tu-freiberg.de>
Subject: [HACKERS] password authentication
This patch adds support for plaintext password authentication. To use
it, you add a line like
host all 0.0.0.0 0.0.0.0 password pg_pwd.conf
to your pg_hba.conf, where 'pg_pwd.conf' is the name of a file containing
the usernames and password hashes in the format of the first two fields
of a Unix /etc/passwd file. (Of course, you can use a specific database
name or IP instead.)
Then, to connect with a password through libpq, you use the PQconnectdb()
function, specifying the "password=" tag in the connect string and also
adding the tag "authtype=password".
I also added a command-line switch '-u' to psql that tells it to prompt
for a username and password and use password authentication.
Subject: [HACKERS] linux/alpha patches
These patches lay the groundwork for a Linux/Alpha port. The port doesn't
actually work unless you tweak the linker to put all the pointers in the
first 32 bits of the address space, but it's at least a start. It
implements the test-and-set instruction in Alpha assembly, and also fixes
a lot of pointer-to-integer conversions, which is probably good anyway.
Subject: [HACKERS] linux/alpha patches
These patches lay the groundwork for a Linux/Alpha port. The port doesn't
actually work unless you tweak the linker to put all the pointers in the
first 32 bits of the address space, but it's at least a start. It
implements the test-and-set instruction in Alpha assembly, and also fixes
a lot of pointer-to-integer conversions, which is probably good anyway.
Subject: [HACKERS] equal column and table name patch
This fixes a bug where selects fail when there is a column with the same
name as the table it's a part of.
Subject: [HACKERS] better access control error messages
This patch replaces the 'no such class or insufficient privilege' with
distinct error messages that tell you whether the table really doesn't
exist or whether access was denied.
Subject: [HACKERS] backend Makefile patch
This patch cleans up backend/Makefile a little bit, and prevents it from
relinking the backend binary when no changes have been made.
Subject: [HACKERS] abort failed transaction patch
This patch allows you to end a transaction that has failed on an error
using the 'ABORT' statement without generating another error message.
(By default you get an error unless you use 'END' to terminate the
transaction, which has already been aborted anyway.)
${DATADIR}. The file is left as pg_geqo.sample, since, unlike
pg_hba.conf, it isn't a required file...but this way ppl know that
its there, and that its where it is required, if they choose to
use it
The following patch to src/backend/libpq/pqpacket.c provides additional
checking for bad packet length data. It was tested with the Linux telnet
client, with netcat using the numbers.txt and by dumping random numbers
into the port.
Patch by: Alvaro Martinez Echevarria <alvaro@lander.es>
The following patches add to the backend a new debugging flag -K which prints
a debug trace of all locking operations on user relations (those with oid
greater than 20000). The code is compiled only if LOCK_MGR_DEBUG is defined,
so the patch should be harmless if not explicitly enabled.
I'm using the code to trace deadlock conditions caused by application queries
using the command "$POSTMASTER -D $PGDATA -o '-d 1 -K 1'.
The patches are for version 6.0 dated 970126.
Patches from: aoki@CS.Berkeley.EDU (Paul M. Aoki)
i gave jolly my btree bulkload code a long, long time ago but never
gave him a bunch of my bugfixes. here's a diff against the 6.0
baseline.
for some reason, this code has slowed down somewhat relative to the
insertion-build code on very small tables. don't know why -- it used
to be within about 10%. anyway, here are some (highly unscientific!)
timings on a dec 3000/300 for synthetic tables with 10k, 100k and
1000k tuples (basically, 1mb, 10mb and 100mb heaps). 'c' means
clustered (pre-sorted) inputs and 'u' means unclustered (randomly
ordered) inputs. the 10k table basically fits in the buffer pool, but
the 100k and 1000k tables don't. as you can see, insertion build is
fine if you've sorted your heaps on your index key or if your heap
fits in core, but is absolutely horrible on unordered data (yes,
that's 7.5 hours to index 100mb of data...) because of the zillions of
random i/os.
if it doesn't work for you for whatever reason, you can always turn it
back off by flipping the FastBuild flag in nbtree.c. i don't have
time to maintain it.
good luck!
baseline code:
time psql -c 'create index c10 on k10 using btree (c int4_ops)' bttest
real 8.6
time psql -c 'create index u10 on k10 using btree (b int4_ops)' bttest
real 9.1
time psql -c 'create index c100 on k100 using btree (c int4_ops)' bttest
real 59.2
time psql -c 'create index u100 on k100 using btree (b int4_ops)' bttest
real 652.4
time psql -c 'create index c1000 on k1000 using btree (c int4_ops)' bttest
real 636.1
time psql -c 'create index u1000 on k1000 using btree (b int4_ops)' bttest
real 26772.9
bulkloading code:
time psql -c 'create index c10 on k10 using btree (c int4_ops)' bttest
real 11.3
time psql -c 'create index u10 on k10 using btree (b int4_ops)' bttest
real 10.4
time psql -c 'create index c100 on k100 using btree (c int4_ops)' bttest
real 59.5
time psql -c 'create index u100 on k100 using btree (b int4_ops)' bttest
real 63.5
time psql -c 'create index c1000 on k1000 using btree (c int4_ops)' bttest
real 636.9
time psql -c 'create index u1000 on k1000 using btree (b int4_ops)' bttest
real 701.0
Essentially, config.h now includes an 'os.h', which is created via
configure by linking a "port.h" file from the port directory to the
include directory.
Going to try to merge backend/port in similar ways
Add a check to configure for strdup
Remove all the '-ltermcap' checks from psql/Makefile
Have {psql,pg_dump}/Makefile modified if strdup doesn't exist on the system
|by neglecting to quote them.
|
|I have made a minor change to pg_dump.c that will fix this.
|
|Dates are dumped and restored OK with pg_dump in V6
|
|We'll still need to fix the dump in both cases if the original dump is from V1.09.
From Keith Parks
Add CFLAGS= @CPPFLAGS@ to Makefile.global and configure so that build is
useful for finding extra header files
Split header files from libraries in build. Doesn't make much sense to
look for a header file in /usr/local/lib, nor to look for a library
in /usr/local/include :)
|Subject: [PATCH] adding SYS_TIME just for fun.
|
|Hi,
|
|Whilst I was playing round with the European dates patch I noticed the sysfunc()
|that allows you to do :-
|
|create table test ( da date);
|insert into test values (SYS_DATE);
|
|and have the current system date inserted.
|
|So I thought it would be nice to have the SYS_TIME facility too.
|
|I've cloned the function and changed a few things and there you have it,
|you can now do:
|
|create table test2 ( ti time);
|insert into test2 values (SYS_TIME);
be #ifdef'd into psql.c itself
From what I can tell, if USE_READLINE is true or false, psql works under
FreeBSD, without configure. Now to test it *again* under sparc_solaris
with configure and see if it works...
history.h doesn't...previously, it was assumed that both existed, or
didn't exist...but this assumption fails on the one sparc_solaris box
that I have access to, and could exist in other circumstances
#if defined(aix)
#define TERMIOS_H_LOCATION <termios.h>
#else
#define TERMIOS_H_LOCATION <sys/termios.h>
#endif
libpq/fe-exec.c modified so that location of termios.h is determined
by whether HAVE_TERMIOS_H is defined or not, in preparation for switch
to configure
The first patch changes the behavior of aclcheck for groups. Currently an user
can access a table only if he has the required permission for ALL the groups
defined for that table. With my patch he can access a table if he has the
permission for ONE of the groups, which seems to me a more useful thing.
If you think this should be the correct behavior of the acl group check feel
free to remove the #ifdef, if not please add a commented line to config.h.
2. IndexScanableOperand now uses match_indexkey_operand
instead of equal_indexkey_var (if we have some index on attribute X
then we shouldn't use it for 'where some_func(X) OP CONST').
/usr/include/limits.h (which quiets the costsize.c warnings)...under
FreeBSD, /usr/include/limits.h *includes* machine/limits.h, while under
Solaris, there is no such things as /usr/include/machine...
Problem with Solaris pointed out by Mark Wahl
1. New flag - BM_JUST_DIRTIED - added for BufferDesc;
2. All data "dirtiers" (WriteBuffer and WriteNoReleaseBuffer)
set this flag (and BM_DIRTY too);
3. All data "flushers" (FlushBuffer, BufferSync and BufferReplace)
turn this flag off just before calling smgr[blind]write/smgrflush
and check this flag after flushing buffer: if it turned ON then
BM_DIRTY will stay ON.
another one in Solaris' port-protos.h.
The following patch will bring inet_aton's prototype into scope for
Ultrix to silence a compilation warning.
If the intention is to have inet_aton's prototype in its own header
filer, the declaration in Solaris' port-protos.h should be removed.
If the declaration in port-protos.h is deemed to be the correct
place, a declaration should be added in Ultrix' port-protos.h
regards
Erik Bertelsen
included after storage/ipc.h like other similar cases that were changed
recently.
This one has popped up during the last few days.
My sources are sup'ed today, 13. jan 1996.
regards
Erik Bertelsen.
Here is a trivial patch to get back the 1.09 behavior; it just removes trailing
newlines before printing the line out with a newline rather than after...
Thomas Lockhart
At least the first two should be fixed before the final release of 6.0.
1) There is a mismatch between the type declared in the catalog for
the input/output attributes of pg_type and the actual type of
values stored in the table. The type of typinput, typoutput,
typsend and typreceive are declared oid (26) while the values are
regproc (24). The error was there also in previous versions but
nobody noticed it until an Assert has been added in ExecEvalVar.
The effect is that it is now impossible to replace the typoutput
of existing data types with new procs.
2) The identd hba fails after the first time because the data read
from the identd socket is not zero-terminated and strlen reports
an incorrect length if the stack contains garbage, which usually
happens after the first connection has been made.
3) The new initdb wants to create itself the data directory. This
implies that the parent directory must be writable by postgres and
this may not always be desirable. A better solution would be to
allow the directory to be created by root and then filled by initdb.
It would also nice to have some reasonable default for PGLIB and
PGDATA like the previous version did. This applies also to the
postmaster executable.
Hi,
counting the empty dummy queries in libpq isn't everything.
If the backend sends an error, the I returns from the dummies
still come. So we must eat them up in any case, not just
returning on the occurence of an E reply.
Until later, Jan
these routines try to use the old pointer casting stuff to get
the connection id, second the notification hash table should
be part of the cliendData. Otherwise, one interpreter might
eat up the notifies for another one.
Please apply the patch below to the current 6.0 tree.
Submitted by: wieck@sapserv.debis.de
code. I have also written a complete complex number package based on this
tutorial; I will submit this as a contribution soon. Is there a particular
format for contributed tar files? I have a C source file, two SQL files, and a
Makefile.
Thomas Lockhart
gmake of the code without interruption.
There's also some tidy-up of the MAXPATHLEN stuff based on the assumption that
all supported platforms have MAXPATHLEN defined in <sys/param.h>.
(The only unknowns for the above are AIX and IRIX5.)
And now - JMP_BUF again. Is it enough, folks ?
Fixed again:
cc1: warnings being treated as errors
exc.c: In function 'ExcRaise':
exc.c:187: warning: passing arg 1 of 'Longjmp' from incompatible pointer type
gmake[3]: *** [exc.o] Error 1
%ud in a printf format strings instead of just %u.
There were three occurances of this in catalog_utils.c,
two in parser.c and one in rewriteSupport.c in the oid
patch that I submitted and was applied. They won't crash
anything, but the error messages will have a 'd' after the
Oid. Annoying, but none are db-threatening.
Sorry about that folks...I'll be more careful in the future...
Darren King
As an example I sent a bug-report on 26 Nov to tell that the fix included
below is necessary to compile pg95-current on Ultrix with Digital's
standard C compiler c89. In fact I think that this fix is needed
for any C compiler sticking very close the standard, see my discussion
in the original bug report.
Erik Bertelsen
cc1: warnings being treated as errors
datum.c: In function `DatumGetSize':
datum.c:57: warning: unsigned value >= 0 is always 1
gmake[3]: *** [datum.o] Error 1
There was:
if (byVal) {
if (len >= 0 && len <= sizeof(Datum)) {
but len has type Size (unsigned int) and so now there is:
if (byVal) {
if (len <= sizeof(Datum)) {
cc1: warnings being treated as errors
exc.c: In function 'ExcRaise':
exc.c:186: warning: passing arg 1 of 'Longjmp' from incompatible pointer type
gmake[3]: *** [exc.o] Error 1
Now we have:
#if defined (JMP_BUF)
longjmp(efp->context, 1);
#else
siglongjmp(efp->context, 1);
#endif
cc1: warnings being treated as errors
transsup.c: In function `TransBlockGetLastTransactionIdStatus':
transsup.c:122: warning: unsigned value >= 0 is always 1
gmake[3]: *** [transsup.o] Error 1
...
(old _bt_compare always returned >= 0 while comparing with P_HIKEY
on root page - it breaks root page when _bt_insertonpg tries insert
new minimal key into root page).
2. Fixed bug concerns "empty" pages: non-rightmost pages with only P_HIKEY
present on it. Such pages appear after vacuum.
as ints and longs. Touches on quite a few function args as
well. Most other files look ok as far as Oids go...still checking
though...
Since Oids are type'd as unsigned ints, they should prolly be used
with the %ud format string in elog and sprintf messages. Not sure
what kind of strangeness that could produce.
Darren King
When an acl item is added or updated the new entry is deleted if it has no
permissions and the acl array is shrinked. This is is done by decrementing
the number of items without updating the corresponding array size.
The array with the incorrect size is later read by pg_aclcheck and the entry
count is used to allocate a new array while the array size is used to copy
the old one. This causes a memory corruption and a backend crash.
This happens only to normal user as the administrator bypasses acl checks.
Massimo Dal Zotto
PQexec handles the possibility of multiple results from one
query by simply submitting an empty query after the first
result and waiting for an 'I' message.
Rules can generate errors with transaction abort after the
first 'C' message was recieved (e.g. if a C-language function
used in a rule calls elog(WARN, ...)). Thus we have to look
for.
Jan(wieck@sapserv.debis.de)
* Wrote max(date) and min(date) aggregates
* Wrote operator "-" for date; date - date yields number of days
difference
* Wrote operator+(date,int) and operator-(date,int); the int is the
number of days. Each operator returns a new date.
By: Tom Tromey <tromey@creche.cygnus.com>
In particular, no more compiled-in default for PGDATA or LIBDIR. Commands
that need them need either invocation options or environment variables.
PGPORT default is hardcoded as 5432, but overrideable with options or
environment variables.
Changes:
* Unique index capability works using the syntax 'create unique
index'.
* Duplicate OID's in the system tables are removed. I put
little scripts called 'duplicate_oids' and 'find_oid' in
include/catalog that help to find and remove duplicate OID's.
I also moved 'unused_oids' from backend/catalog to
include/catalog, since it has to be in the same directory
as the include files in order to work.
* The backend tries converting the name of a function or aggregate
to all lowercase if the original name given doesn't work (mostly
for compatibility with ODBC).
* You can 'SELECT NULL' to your heart's content.
* I put my _bt_updateitem fix in instead, which uses
_bt_insertonpg so that even if the new key is so big that
the page has to be split, everything still works.
* All literal references to system catalog OID's have been
replaced with references to define'd constants from the catalog
header files.
* I added a couple of node copy functions. I think this was a
preliminary attempt to get rules to work.
to be sleazy and reach into other subsystems' directories. First entry in
this directory is the PG_VERSION file interface, which must be used by the
backend and also the pg_version program (which is used by initdb).
|After sending my previous changes I found one more thing in Makefile.global.
|Any tests should be done after including Makefile.custom or else there
|won't be anything, particularly the port name, to base the tests on.
|
-----
more changes to makefile.global from D'Arcy
following is the patch to libpq's large object interface that
removes the requirement to include fmgr.h into fe-lobj.c.
The large object interface now ask's the backend to tell the
OID's of all the required functions in pg_proc.
From: wieck@sapserv.debis.de (Jan Wieck)
>From the create_aggregate man page...
"The arguments to state-transition-function-1 must be
(stype1,basetype), and its return value must be stype1."
create aggregate MIN (sfunc1 = int2smaller,
basetype = int2,
stype1 = int2);
will fail becase int2smaller and int2larger are in pg_proc
as returning an int4. Can't happen since both args have to
be int2.
From: Darren King <aixssd!ceodev!darrenk@abs.net>
other platforms). If I do the standard make + make install the shared library
is not linked with the the libpq library and when I try to load it in the
standard Tcl or Tk shell I get a lot of unresolved symbols. The bug doesn't
affect pgtclsh because it is linked also with libpq. The problem exists only
when using the dynamic load feature of Tcl7.5.
From: Massimo Dal Zotto <dz@cs.unitn.it>
I'm able to get through a 'make' of the backend with no errors except
the occasional 'might not be initialized error', which is nothing major,
just annoying.
Have a few patches from D'Arcy to incorporate, but am waiting until I can
get a clean compile first, which I'm hoping to have before bed, or sometime
tomorrow.
Note. all include files that have been hit so far have had extraneous
include files cleaned out and are reduced to...the lowest common
"include file", based on 'cc -Wall -I. test.c', where test.c is:
#include "postgres.h"
#include "<top of branches>" (ie. top of branches this time was utils/fcache2.h)
*should* be intelligent enough that:
#if defined(__FreeBSD__) works, where __FreeBSD__ is actually defined
by the compiler itself.
Makefile.global used to have -DPORTNAME_<port> -D<port> as part of the flags
for gcc while all occurances of PORTNAME_<port> slowly get removed from
the source tree...
Adds:
-lAttributes
Returns another format of the results attribute list. Per
attribute a sublist of {{attname} atttype attlen} is
returned and an empty string if no attributes where
received.
-numAttrs
Returns the number of attributes in the result.
I found another bug in btree index. Looking at the code it seems that NULL
keys are never used to build or scan a btree index (see the explain commands
in the example). However this is not the case when a null key is retrieved
in an outer loop of a join select and used in an index scan of an inner loop.
This bug causes at least three kinds of problems:
1) the backend crashes when it tries to compare a text string with a null.
2) it is not possible to find tuples with null keys in a join.
3) null is considered equal to 0 when the datum is passed by value, see
the last query.
Submitted by: Massimo Dal Zotto <dz@cs.unitn.it>
It's bug in nodeAgg.c on lines 241, 242:
null_array = malloc(nagg);
for (i=0;i<nagg;i++)
null_array[i] = 'n';
oneTuple = heap_formtuple(tupType, tupValue, null_array);
- your query has not only aggregates but also 'group by-ed' fields and so
null_array should contain tupType->natts elements (tupType->natts > nagg in
your case).
Patch follows and it's very simple.
VAdim
My guess is that the thing had bugs, and the pfree was commented out.
The thing is probabally free'ed anyway at the end, so it was not a bad
thing.
If it does cause a bug, it will generate an error when hit, so I say
unless someone else knows, let's remove it and run the regression test.
-Bruce
|
| 1. Set default variables
| 2. Include Makefile.custom to override defaults
| 3. Set CFLAGS, etc. with variables
|
|This fixes the problem of Makefile.custom changes not taking effect.
Submitted by: D'Arcy Cain
2. Reap unused tuples too
3. Reap empty pages
4. Check if a page is initialized, initialize it if not
and reap it
5. Binary search in list of reapped pages/tids to check
is the heap' tid pointed by a index' tuple on this list
(it's mu-u-uch faster)
cache. I found if I manually added a line to flush the whole relation
cache, the assert error disappeared. Looking through the code, I found
that the relation cache is flushed at the end of each query if the
reference count is zero for the relation. However, printf's showed that
the rd_relcnt(reference count) for the accessed query was not returning
to zero after each query.
It turns out the parser was doing a heap_ropen in parser/analyze.c to
get information about the table's columns, but was not doing a
heap_close.
This was causing the query after the ALTER TABLE ADD to see the old
table structure, and the executor's assert was reporting the problem.
correct way to do this. Theoretically you could have a NULL
pointer that isn't represented internally as all 0 bits. This
guarantees that it convert correctly.
Submitted by: darcy@druid.com (D'Arcy J.M. Cain)
NAMEDATALEN
OIDDATALEN
EUROPEAN_DATES
HBA
DEADLOCK_TIMEOUT
OPENLINK_PATCHES
NULL_PATCH
ARRAY_PATCH
Attempting to document and centralize as many of the "defines" as possible...
kinda useless to have defines if nobody knows they exist, eh?
function so I am going to assume that it is such a good idea that no
one sees any point in discussing it. :-) I have made two changes -
I have merged this into pgtclCmds.c and I have taken out any code for
updating tuples after the loop body runs. See comments for discussion
of this.
I have also fixed up the error checking stuff so that break, continue
and syntax errors have the expected result.
Submitted by: D'Arcy Cain
my postmaster 1.07.
It's really simple, the loop dealing with all sockets
can't handle more than one ready socket :-)
A simple logic error dealing with lists.
OR IS THERE ANY REASON FOR SETTING curr TO 0?
Submitted by: Carsten Heyl <Heyl@nads.de>
with some versions of sh, and a bug in the master make file that
causes it to issue the message "postgres has been built" at the wrong
time.
Submitted by: bryanh@giraffe.netgate.net (Bryan Henderson)
To: Postgres95-development <pg95-dev@ki.net>
Subject: [PG95-DEV] postgres.h patch
This removes the parameters from the func_ptr prototype in postgres.h in
2.0.
To: Postgres95-development <pg95-dev@ki.net>
Subject: [PG95-DEV] Makefile.global patch
Can we apply the following patch to make EUROPEAN_DATES consistent with
other parameters?
way one creates a database system. Parts that were in "make install"
are not either in "make all" or initdb. Nothing goes in the PGDATA
directory besides user data. Creating multiple database systems is
easier.
In addition to applying the patch, it is necessary to move the file
libpq/pg_hba to backend/libpq/pg_hba.sample.
Submitted by: Bryan Henderson <bryanh@giraffe.netgate.net>
--- src/backend/storage/file/fd.c Thu Sep 12 17:23:38 1996
***************
*** 262,268 ****
Delete(file);
/* save the seek position */
! fileP->seekPos = lseek(fileP->fd, 0L, SEEK_CUR);
Assert( fileP->seekPos != -1);
/* if we have written to the file, sync it */
--- 262,268 ----
Delete(file);
/* save the seek position */
! fileP->seekPos = (long) lseek(fileP->fd, 0L, SEEK_CUR);
Assert( fileP->seekPos != -1);
/* if we have written to the file, sync it */
Submitted by: Randy Terbush <randy@zyzzyva.com>
- Added the header access/heapam.h.
- Changed all instances of "length" to "data_length" to quiet
the compiler.
- initialized a few variables. The compiler couldn't see that
the code guaranteed that these would be initialized before
being dereferenced. If anyone wants to check my work follow
the usage of these variables and make sure that this true
and wasn't actually a bug in the original code.
- added a missing break statement to a default case. This
was a benign error but bad style.
- layed out heap_sysattrlen differently. I think this way
makes the structure of the code crystal clear. There should
be no actual difference in the actual behaviour of the code.
Submitted by: darcy@druid.druid.com (D'Arcy J.M. Cain)
NOTE: Makefile.custom is commented out, since it isn't there by default.
If you read the section telling you about it to know to create it,
you can uncomment it while you are there ...
current state of development...namely, we are on 2.0
NOTE:
BTW, the is also a check in postmaster which won't let you use an older
version of the database by checking the version number. The version number
of a database is in data/PG_VERSION (a plain ASCII file).
- Andrew
attributes as tcl arrays. The previous code had problems with some chars
used as delimiter by Tcl. The new code should be more robust.
By: Massimo Dal Zotto <dz@cs.unitn.it>
Async notifies received while a backend is in the middle of a begin/end
transaction block are lost by libpq when the final end command is issued.
The bug is in the routine PQexec of libpq. The routine throws away any
message from the backend when a message of type 'C' is received. This
type of message is sent when the result of a portal query command with
no tuples is returned. Unfortunately this is the case of the end command.
As all async notification are sent only when the transaction is finished,
if they are received in the middle of a transaction they are lost in the
libpq library. I added some tracing code to PQexec and this is the output:
Submitted by: Massimo Dal Zotto <dz@cs.unitn.it>
execute an sql function containing an utility command (create, notify, ...).
The bug is part in the planner, which returns a number of plans different
than the number of commands if there are utility commands in the query, and
in part in the function executor which assumes that all commands are normal
query commands and causes a SIGSEGV trying to execute commands without plan.
Submitted by: Massimo Dal Zotto <dz@cs.unitn.it>
|
|Here's a patch for Version 2 only. It just adds an Assert to catch some
|inconsistencies in the catalog classes.
|
|--
|Bryan Henderson Phone 408-227-6803
|San Jose, California
|
The problem is that the function arguments are not considered as possible key
candidates for index scan and so only a sequential scan is possible inside
the body of a function. I have therefore made some patches to the optimizer
so that indices are now used also by functions. I have also moved the plan
debug message from pg_eval to pg_plan so that it is printed also for plans
genereated for function execution. I had also to add an index rescan to the
executor because it ignored the parameters set in the execution state, they
were flagged as runtime variables in ExecInitIndexScan but then never used
by the executor so that the scan were always done with any key=1. Very odd.
This means that an index rescan is now done twice for each function execution
which uses an index, the first time when the index scan is initialized and
the second when the actual function arguments are finally available for the
execution. I don't know what is the cost of an double index scan but I
suppose it is anyway less than the cost of a full sequential scan, at leat
for large tables. This is my patch, you must also add -DINDEXSCAN_PATCH in
Makefile.global to enable the changes.
Submitted by: Massimo Dal Zotto <dz@cs.unitn.it>
The comparison routines for text and char data type give incorrect results
if the input data contains characters greater than 127. As these routines
perform the comparison using signed char variables all character codes
greater than 127 are interpreted as less than 0. These codes are used to
encode the iso8859 char sets.
The other text-like data types seem to work as expected as they use unsigned
chars in comparisons.
Submitted by: Massimo Dal Zotto <dz@cs.unitn.it>
tree, instead of having include files all over the place...
Immediate goal...a 'config.h' file so that we can make #ifdef's
being used throughout the code more a rarity as far as porting
is concerned
conditions are always met. The patch can be applied to any version
of Postgres95 from 1.02 to 1.05. After applying the patch, queries
using indices on bpchar and varchar fields should (hopefully ;-) )
always return the same tuple set regardless to the fact whether
indices are used or not.
Submitted by: Gerhard Reithofer <tbr_laa@AON.AT>
In a catalog class that has a "name" type attribute, UPDATEing of an
instance of that class may destroy all of the attributes of that
instance that are stored as or after the "name" attribute.
This is caused by the alignment value of the "name" type being set to
"double" in Class pg_type, but "integer" in Class pg_attribute.
Postgres constructs a tuple using double alignment, but interprets it
using integer alignment.
The fix is to change the alignment to integer in pg_type.
Note that this corrects the problem for new Postgres systems. Existing
databases already contain the error and it can't easily be repaired because
this very bug prevents updating the class that contains it.
--
Bryan Henderson Phone 408-227-6803
San Jose, California
It adds a WITH OIDS option to the copy command, which allows
dumping and loading of oids.
If a copy command tried to load in an oid that is greater than
its current system max oid, the system max oid is incremented. No
checking is done to see if other backends are running and have cached
oids.
pg_dump as its first step when using the -o (oid) option, will
copy in a dummy row to set the system max oid value so as rows are
loaded in, they are certain to be lower than the system oid.
pg_dump now creates indexes at the end to speed loading
Submitted by: Bruce Momjian <maillist@candle.pha.pa.us>
This presumably corrects a problem of initdb failing on systems that have
an awk that is sensitive to this.
--
Bryan Henderson Phone 408-227-6803
San Jose, California
|record, by a small patch to libpq++? At least until the
|feature that will allow dumped oid's to be re-loaded into
|a database becomes available, I need access to the oids
|of newly created records... To this end, I have written a
|three-line wrapper for the PQoidStatus function in libpq and
|named this wrapper OidStatus() (I'd appreciate suggestions for
|a name that would better fit into the general naming scheme).
|
|Regards,
|
|Ernst
|
When you try to do any UPDATE of the catalog class pg_class, such as
to change ownership of a class, the backend crashes.
This is really two serial bugs: 1) there is a hardcoded copy of the
schema of pg_class in the postgres program, and it doesn't match the
actual class that initdb creates in the database; 2) Parts of postgres
determine whether to pass an attribute value by value or by reference
based on the attbyval attribute of the attribute in class
pg_attribute. Other parts of postgres have it hardcoded. For the
relacl[] attribute in class pg_class, attbyval does not match the
hardcoded expectation.
The fix is to correct the hardcoded schema for pg_attribute and to
change the fetchatt macro so it ignores attbyval for all variable
length attributes. The fix also adds a bunch of logic documentation and
extends genbki.sh so it allows source files to contain such documentation.
--
Bryan Henderson Phone 408-227-6803
San Jose, California
---
below my signature, there are a coupls of diffs and files in a shell
archive, which were needed to build postgres95 1.02 on Siemens Nixdorfs
MIPS based SINIX systems. Except for the compiler switches "-W0" and
"-LD-Blargedynsym" these diffs should also apply for other SVR4 based
systems. The changes in "Makefile.global" and "genbki.sh" can probably
be ignored (I needed gawk, to make the script run).
There is one bugfix thou. In "src/backend/parser/sysfunc.c" the
function in this file didn't honor the EUROPEAN_DATES ifdef.
---
Submitted by: Frank Ridderbusch <ridderbusch.pad@sni.de>
Here's a couple more small fixes that I've made to make my runtime
checker happy with the code. More along the lines of those that
I sent in the past, ie, a pointer to an array != the name of
an array. The last patch is that I mailed about yesterday -- I got
two replies of "do it", so it's done. As far as I can tell, however,
the function in question is never called by pg95, so either way
it can't hurt...
From: "Kurt J. Lidl" <lidl@va.pubnix.com>
When you connect to a database with PQsetdb, as with psql, depending on
how your uninitialized variables are set, you can get a failure with a
"There is no connection to the backend" message.
The fix is to move a call to PQexec() from inside connectDB() to
PQsetdb() after connectDB() returns to PQsetdb(). That way a connection
doesn't have to be already established in order to establish it!
From: bryanh@giraffe.netgate.net (Bryan Henderson)
|
|This patch fixes a backend crash that happens sometimes when you try to
|join on a field that contains NULL in some rows. Postgres tries to
|compute a hash value of the field you're joining on, but when the field
|is NULL, the pointer it thinks is pointing to the data is really just
|pointing to random memory. This forces the hash value of NULL to be 0.
|
|It seems that nothing matches NULL on joins, even other NULL's (with or
|without this patch). Is that what's supposed to happen?
|
CLUSTER command couldn't rename correctly the new created heap relation.
The table base name resulted in some "temp_XXXX" instead of the correct
base name.
Submitted by: Dirk Koeser <koeser@informatik.uni-rostock.de>
Postgres is not able to cluster a relation on which an rtree index is
defined. Postmaster gives the following error message:
Too Large Allocation Request("!(0 < (size) && (size) <= (0xfffffff)):size=0
[0x0]", File:"/export/home/postgres/src/backend/utils/mmgr/mcxt.c", Line: 220)
!(0 <(size) && (size) <= (0xfffffff)) (0) [No such file or directory]
Submitted by: Dirk Koeser <koeser@informatik.uni-rostock.de>
|Here is a fix for the psql alignment problem. It turns out that libpq
|was trying to determine if the column contained only numeric values so
|it could right justify it. The 'e' values were taked as exponient
|values and all columns were considered numeric.
|
|The patch excludes 'e' and 'E' as being valid first-column numeric
|values.
|
Submitted by: Bruce...
pg_dump and load to 2.0. I haven't gotten any feedback on whether
people want it, so I am submitting it for others to decide. I would
recommend an install in 1.02.1.
I had said that the 2.0 pg_dump could dump a 1.02.1 database, but I was
wrong. The copy is actually performed by the backend, and the 2.0
database will not be able to read 1.02.1 databases because of the new
system columns.
This patch does several things. It copies nulls out as \N, so they can
be distinguished from '' strings. It fixes a problem where backslashes
in the input stream were not output as double-backslashes. Without this
patch, backslashes copied out were deleted upon input, or interpreted as
special characters. Third, input is now terminated by backslash-period.
This can not be part of a normal input stream.
I tested this by creating a database with all sorts of nulls, backslash,
and period fields and dumped the database and reloaded into a new
database and compared them.
Submitted by: Bruce
pg_dump and load to 2.0. I haven't gotten any feedback on whether
people want it, so I am submitting it for others to decide. I would
recommend an install in 1.02.1.
I had said that the 2.0 pg_dump could dump a 1.02.1 database, but I was
wrong. The copy is actually performed by the backend, and the 2.0
database will not be able to read 1.02.1 databases because of the new
system columns.
This patch does several things. It copies nulls out as \N, so they can
be distinguished from '' strings. It fixes a problem where backslashes
in the input stream were not output as double-backslashes. Without this
patch, backslashes copied out were deleted upon input, or interpreted as
special characters. Third, input is now terminated by backslash-period.
This can not be part of a normal input stream.
I tested this by creating a database with all sorts of nulls, backslash,
and period fields and dumped the database and reloaded into a new
database and compared them.
Submitted by: Bruce
and found out that one of the patches is a show stopper for
compiling under a strict ansi package.
Please make sure the following fix makes it into the 1.02.1
release...
Thanks.
-Kurt
|We're all too familiar with psql's "no response from backend" message.
|Users can't tell what this means, and psql continues prompting for
|commands after it even though the backend is dead and no commands can
|succeed. It eventually dies on a signal when the dead socket fills
|up. I extended the message to offer a better explanation and made
|psql exit when it finds the backend is dead.
|
|I also added a short message and newline when the user does a ctl-D so
|it doesn't mess up the terminal display.
|
|
Submitted by: Bryan Henderson <bryanh@giraffe.netgate.net>
don't indicate that the libpq.a library is a dependency of all the /bin
programs. So if the library changes, the /bin programs don't get remade.
Submitted by: Bryan Henderson <bryanh@giraffe.netgate.net>
directory. The code that looks for the pg_hba file doesn't use it, though,
so the postmaster uses the wrong pg_hba file. Also, when the postmaster
looks in one directory and the user thinks it is looking in another
directory, the error messages don't give enough information to solve the
problem. I extended the error message for this.
Submitted by: Bryan Henderson <bryanh@giraffe.netgate.net>
does 2 things:
1) Make it hard to not notice the make failed. (As you recall, someone on
the mailing list had this problem. I've had it to some extent myself).
The 1.02 make files continue with the next subdirectory when a make
in a subdirectory fails. The patch makes the make stop in the
conventional way when a submake fails. It also adds a reassuring message
when the make succeeds and adds a note to the INSTALL file to expect it.
2) Include loader flags on all invocations of the linker.
The 1.02 make files omit the $(LDFLAGS) on some of the linker invocations.
On my system, I need one of those flags just to make it invoke the proper
version of the compiler/linker, so LDFLAGS has to be everywhere.
Submitted by: Bryan Henderson <bryanh@giraffe.netgate.net>
Attached is a patch to allow libpq to determine if a field is null.
This is needed because text fields will return a PQgetlength() of 0
whether it is '' or NULL. There is even a comment in the source noting
the fact.
I have changed the value of the 'len' field for NULL result fields. If
the field is null, the len is set to -1 (NULL_LEN). I have changed
PQgetlength() to return a 0 length for both '' and NULL. A new function
PQgetisnull() returns true or false for NULL.
The only risk is to applications that do not use the suggested
PQgetlength() call, but read the result 'len' field directly.
As this is not recommended, I think we are safe here.
A separate documentation patch will be sent.
Submitted by: Bruce Momjian <maillist@candle.pha.pa.us>
Here's a small patch that my run-time checker whines about
incessantly. The justification for the patch is along the
lines of passing a NULL is allowed if you have an
arguement that is a *POINTER* to something, but if
the arguement is an array reference, it's not really
a "pointer", so it can't be NULL.
If you question this, I refer you to
<URL:http://www.va.pubnix.com/staff/djm/lore/arrays-are-not-pointers>
Anyways, here's the patch:
-Kurt
Submitted by: "Kurt J. Lidl" <lidl@va.pubnix.com>
This patch forces postgres95 to assume any floating-point value is a
float8. It removes the requirement that you cast all floating-point
constants to float8.
We can remove alot of casts in the regression test after we are sure
this works.
If I have missed anything, would someone let me know. I have tested
inserts of floating-point values into float8 fields, and it worked well.
Casting the number to float4 showed the same precision loss as previous
uncast values showed.
Submitted by: Bruce Momjian <maillist@candle.pha.pa.us>
There is a support routine in the standard 4.4BSD C library
called "err()". There is also a utility routine in
.../src/backend/bootstrap/bootstrap.c
with the same name.
Here's a patch that renames the pg95 routine to something a little
more sane. As a bonus, one more bit of system-specific code leaves
the system...
Submitted by: "Kurt J. Lidl" <lidl@va.pubnix.com>
Also, I think that an extra source of noise in the diff of regress.out and
expected.out is caused by not substituting the shared library file
extension in the regression.input file (much like the paths and the
usernames are sub'ed). This seems to be fixed with the following patches
to regression.input and the Makefile... If I'm off base here, please tell!
Submitted by: Wayde Nie <niew@phoenix.cis.mcmaster.ca>
I've enclosed two patches. The first affects Solaris compilability. The
bug stems from netdb.h (where MAXHOSTNAMELEN is defined on a stock
system). If the user has installed the header files from BIND 4.9.x,
there will be no definition of MAXHOSTNAMELEN. The patch will, if all
else fails, try to include <arpa/nameser.h> and set MAXHOSTNAMELEN to
MAXDNAME, which is 256 (just like MAXHOSTNAMELEN on a stock system).
The second patch adds aliases for "ISNULL" to "IS NULL" and likewise for
"NOTNULL" to "IS NOT NULL". I have not removed the postgres specific
ISNULL and NOTNULL. I noticed this on the TODO list, and figured it would
be easy to remove.
The full semantics are:
[ expression IS NULL ]
[ expression IS NOT NULL ]
--Jason
Submitted by: Jason Wright <jason@oozoo.vnet.net>
Previously Postgres95 wouldn't accept 'order by' clauses with fields
referred to as '<table>.<field>', e.g.:
select t1.field1, t2.field2 from table1 t1, table2 t2
order by t2.field2;
This syntax is required by the ODBC SQL spec.
Submitted by: Dan McGuirk <mcguirk@indirect.com>
While a normal SELECT statement can contain a GROUP BY clause, a cursor
declaration cannot. This was not the case in PG-1.0. Was there a good
reason why this was changed? Are cursors being phased out? Is there any way
to get data with just a SELECT (and without a DECLARE CURSOR ...)?
The patch below seems to fix things. If anyone can see a problem with it,
please let me know. Thanks.
Submitted by: David Smith <dasmith@perseus.tufts.edu>
Here are a few minor fixes to Postgres95. Mostly I have added const
to some of the char pointers. There was also a missing header file
and a place where it looks like "==" was used when "=" was meant.
I also changed some variables from Pfin and Pfout tp pfin and pfout
because the latter shadow global variables and that just seems like
an unsafe practice which I like to avoid.
Submitted by: "D'Arcy J.M. Cain" <darcy@druid.druid.com>
Someone asked me if the bpchar type could be extended to do
case-insensitive regular expression searches.
Submitted by: "Alistair G. Crooks" <azcb0@juts.ccc.amdahl.com>
Originally, I thought the problem was caused by a function that gets
called as a normal function where we want to return a value, and as a
signal handler where we need to have it accept a parameter (the signal
number) and it returns nothing, I was going to case the function name in
the signal call as (void (*)(int)).
Looking at all the source, it turns out this function only gets used as
a signal handler, so I set an int parameter and return void.
I have removed the Linux defines because they are not needed. BSD let
this sloppiness slide. Linux gave a compile error.
Submitted by: Bruce Momjian <maillist@candle.pha.pa.us>
In postgres95/src/backend/nodes/readfuncs, lines 1188 and 1189,
local_node->relname is taken to point to a NameType, while its
defined as a pointer to char. Both the casting to Name and the
call of namestrcpy should, IMHO, be changed appropriately (first
patch).
As far as I could see from the Linux signal header file,
a signal handler is declared as
typedef void (*__sighandler_t)(int);
Few changes to postgres95/src/backend/storage/lmgr/proc.c seem
appropriate to comply with this.
Finally, postgres95/src/bin/pg_version/pg_version.c defines
a function GetDataHome (by default, returning an integer)
and returns NULL in the function, which isn't an integer...
Submitted by: ernst.molitor@uni-bonn.de
updates the psql.1 manual page for \ options
add row count and ties it to the header option
updated manual pages and comment for above change
got \? to display in one screen-full (almost, \? scrolls off top)
moved \r to \E, and \z to \r (for historical reasons with monitor)
small code alignment cleanup
Submitted by: Bruce Momjian <maillist@candle.pha.pa.us>
> INDEXED searches in some cases DO NOT WORK.
> Although simple search expressions (i.e. with a constant value on
> the right side of an operator) work, performing a join (by putting
> a field of some other table on the right side of an operator) produces
> empty output.
> WITHOUT indices, everything works fine.
>
submitted by: "Vadim B. Mikheev" <root@ais.sable.krasnoyarsk.su>
case where the attribute length is variable (stored as -1). Previously,
you'd get output that looked like:
CREATE TABLE foo (bar varchar(-1));
Monitor and psql don't like this at all :). Here is a fix:
Submitted by: Adam Sussman <myddryn@vidya.com>
Kerberos is being used (attempt to free static memory).
The error was caused by a confusing doublespeak of fe_getauthname():
Returns a pointer to static memory, if you authenticate via Kerberos,
a pointer to dynamic memory otherwise.
Submitted by: Erich Stamberger <eberger@gewi.kfunigraz.ac.at>
of my (proff) patch. This is the rest of it, with a few, mainly aesthetic
changes. I've removed a lot of redundency from the original code,
added support for the new PQprint() routines in libpq, expanded tables,
and a few generally nifty ways of massaging data in and out of the
backend. Still needs some good stress testing.
compatibility. There isn't much difference here against my previous
PQprint() code, except that you can add optional arguments to the
<table args> in html.
Most of the changes in here look to b epurely cosmetic, and don't
affect anything...
...and some stuff is completely questionable...in that I may have reversed
some of the stuf fwe already had :(
- src/backend/tcop/*
- cosmetic changes to OPENLINK patches
- src/backend/storage/*
- more changes, mostly cosmetic
- src/backend/ports/*
- merge in patches for aix and i386_solaris
>
> We did some testing and found that if we name the table 'Inv' with
> anything appended to it, the table does not appear in the '\d' table list.
> It appears to be the capital I as a table named 'invItemsL' is created
> and displayed properly.
>
- submitted by: Jason Wright <jason@shiloh.vnet.net>
The updating of array fields is broken in Postgres95-1.01, An array can
be only replaced with a new array but not have some elements modified.
This is caused by two bugs in the parser and in the array utilities.
Furthermore it is not possible to update array with a base type of
variable length.
- submitted by: Massimo Dal Zotto <dz@cs.unitn.it>
I have written some patches which add support for NULLs to Postgres95.
In fact support for NULLs was already present in postgres, but it had been
disabled because not completely debugged, I believe. My patches simply add
some checks here and there. To enable the new code you must add -DNULL_PATCH
to CFLAGS in Makefile.global. After recompiling you can do things like:
insert into a (x, y) values (1, NULL);
update a set x = NULL where x = 0;
You can't still use a "where x=NULL" clause, you must use ISNULL instead.
This could probably be an easy fix to do.
Submitted by: Massimo Dal Zotto <dz@cs.unitn.it>
Select queries with an isnull or notnull clause, like "select * where
somefield isnull", crash the backend if the table has at least one index.
If the indices are deleted the queries work again. Also the explain
command fail in the same way.
The is caused by a bug in subroutine of the optimizer which doesn't check
null values in the clauses.
Submitted by: Massimo Dal Zotto <dz@cs.unitn.it>
This is a patch to prevent an endless loop occuring in the Postgres backend
when a 'warning' error condition generates another warning error contition
in the handler code.
Submitted by: Chris Dunlop, <chris@onthe.net.au>
It is not possible to define attributes as arrays of date or time, the
type _time and _date are not defined.
Submitted by: Massimo Dal Zotto <dz@cs.unitn.it>
The type _char16 (array of char16) is incorrectly defined as array of name
and values longer than 16 chars are stored as names and not truncated to 16
bytes as they should be.
Submitted by: Massimo Dal Zotto <dz@cs.unitn.it>
before (plus some optimisations/bug fixes et al). I've included a small
demo transcript below. Note that all of of the display
functionality/intelligence you see here, can be had merely by calling
the new LIBPQ PQprint() routine with the appropriate arguments/options,
including the HTML3 output guff.
submitted by: Julian Assange <proff@suburbia.net>
varchar.diff
------------
This patch was necessary for the OpenLink Postgres Database Agent.
I think this fixes a bug anyway.
The following query demonstrates this bug:
create table foo (bar varchar);
insert into foo values (''); -- no problem
select * from foo where bar = ''; -- fails
causes segmentation fault.
Thanks to: Salvador Ortiz Garcia, Robert Patrick, Paul 'Shag' Walmsley,
and James Cooper for finding and fixing the problem.
Bruce Momjian