Use exclusive states for parsing quoted strings.
Implement extended comments ("/* ... */") using exclusive states.
Modify definitions of operators to remove some restrictions on characters
and character order.
This commit is contained in:
Thomas G. Lockhart
parent
1370a08746
commit
559a28747a
|
|
@ -8,7 +8,7 @@
|
|||
*
|
||||
*
|
||||
* IDENTIFICATION
|
||||
* $Header: /cvsroot/pgsql/src/backend/parser/scan.l,v 1.14 1997/08/20 01:50:06 vadim Exp $
|
||||
* $Header: /cvsroot/pgsql/src/backend/parser/scan.l,v 1.15 1997/09/01 05:51:52 thomas Exp $
|
||||
*
|
||||
*-------------------------------------------------------------------------
|
||||
*/
|
||||
|
|
@ -59,7 +59,42 @@ void unput(char);
|
|||
#endif /* FLEX_SCANNER */
|
||||
|
||||
extern YYSTYPE yylval;
|
||||
|
||||
int llen;
|
||||
char literal[MAX_PARSE_BUFFER];
|
||||
|
||||
%}
|
||||
/* OK, here is a short description of lex/flex rules behavior.
|
||||
* The longest pattern which matches an input string is always chosen.
|
||||
* For equal-length patterns, the first occurring in the rules list is chosen.
|
||||
* INITIAL is the starting condition, to which all non-conditional rules apply.
|
||||
* <xc> is an exclusive condition to allow embedded C-style comments.
|
||||
* When in an exclusive condition, only those rules defined for that condition apply.
|
||||
* So, when in condition <xc>, only strings which would terminate the "extended comment"
|
||||
* trigger any action other than "ignore".
|
||||
* The "extended comment" syntax closely resembles allowable operator syntax.
|
||||
* Therefore, be sure to match _any_ candidate comment, including those with appended
|
||||
* operator-like symbols. - thomas 1997-07-14
|
||||
*/
|
||||
|
||||
/* define an exclusive condition to allow extended C-style comments - tgl 1997-07-12 */
|
||||
%x xc
|
||||
/* define an exclusive condition for quoted strings - tgl 1997-07-30 */
|
||||
%x xq
|
||||
|
||||
/* We used to allow double-quoted strings, but SQL doesn't so we won't either */
|
||||
quote '
|
||||
xqstart {quote}
|
||||
xqstop {quote}
|
||||
xqdouble {quote}{quote}
|
||||
xqinside [^\']*
|
||||
xqliteral [\\].
|
||||
|
||||
xcline [\/][\*].*[\*][\/]{space}*\n*
|
||||
xcstart [\/][\*]{op_and_self}*
|
||||
xcstop {op_and_self}*[\*][\/]({space}*|\n)
|
||||
xcinside [^*]*
|
||||
xcstar [^/]
|
||||
|
||||
digit [0-9]
|
||||
letter [_A-Za-z]
|
||||
|
|
@ -69,15 +104,15 @@ sysfunc SYS_{letter}{letter_or_digit}*
|
|||
|
||||
identifier {letter}{letter_or_digit}*
|
||||
|
||||
self [,()\[\].;$\:\+\-\*\/\<\>\=\|]
|
||||
op_and_self [\~\!\@\#\%\^\&\|\`\?\$\:\+\-\*\/\<\>\=]
|
||||
op_and_self2 [\~\!\@\#\%\^\&\|\`\?\$\:\*\/\<\>\=]
|
||||
op_only [\~\!\@\#\%\^\&\`\?]
|
||||
typecast "::"
|
||||
|
||||
operator ({op_and_self}{op_and_self2}+)|{op_only}+
|
||||
/* we used to allow double-quoted strings, but SQL doesn't */
|
||||
/* so we won't either*/
|
||||
quote '
|
||||
self [,()\[\].;$\:\+\-\*\/\<\>\=\|]
|
||||
selfm {self}[\-][\.0-9]
|
||||
|
||||
op_and_self [\~\!\@\#\%\^\&\|\`\?\$\:\+\-\*\/\<\>\=]
|
||||
|
||||
operator {op_and_self}+
|
||||
operatorm {op_and_self}+[\-][\.0-9]
|
||||
|
||||
integer -?{digit}+
|
||||
real -?{digit}+\.{digit}+([Ee][-+]?{digit}+)?
|
||||
|
|
@ -97,10 +132,57 @@ other .
|
|||
|
||||
{comment} { /* ignore */ }
|
||||
|
||||
"::" { return TYPECAST; }
|
||||
/* allow extended comments using C-style delimiters - tgl 1997-07-12 */
|
||||
{xcline} { /* ignore */ }
|
||||
|
||||
{self} { return (yytext[0]); }
|
||||
<xc>{xcstar} |
|
||||
{xcstart} { BEGIN(xc); }
|
||||
|
||||
<xc>{xcstop} { BEGIN(INITIAL); }
|
||||
|
||||
<xc>{xcinside} { /* ignore */ }
|
||||
|
||||
{xqstart} {
|
||||
BEGIN(xq);
|
||||
llen = 0;
|
||||
*literal = '\0';
|
||||
}
|
||||
<xq>{xqstop} {
|
||||
BEGIN(INITIAL);
|
||||
yylval.str = pstrdup(scanstr(literal));
|
||||
return (SCONST);
|
||||
}
|
||||
<xq>{xqdouble} |
|
||||
<xq>{xqinside} {
|
||||
if ((llen+yyleng) > (MAX_PARSE_BUFFER - 1)) {
|
||||
elog(WARN,"quoted string parse buffer of %d chars exceeded",MAX_PARSE_BUFFER);
|
||||
/* not reached */
|
||||
}
|
||||
memcpy(literal+llen, yytext, yyleng+1);
|
||||
llen += yyleng;
|
||||
}
|
||||
<xq>{xqliteral} {
|
||||
if ((llen+yyleng-1) > (MAX_PARSE_BUFFER - 1)) {
|
||||
elog(WARN,"quoted string parse buffer of %d chars exceeded",MAX_PARSE_BUFFER);
|
||||
/* not reached */
|
||||
}
|
||||
memcpy(literal+llen, yytext+1, yyleng);
|
||||
llen += yyleng-1;
|
||||
}
|
||||
|
||||
{typecast} { return TYPECAST; }
|
||||
|
||||
{selfm} {
|
||||
yyless(yyleng-2);
|
||||
return (yytext[0]);
|
||||
}
|
||||
{self} { return (yytext[0]); }
|
||||
|
||||
{operatorm} {
|
||||
yyless(yyleng-2);
|
||||
yylval.str = pstrdup((char*)yytext);
|
||||
return (Op);
|
||||
}
|
||||
{operator} {
|
||||
if (strcmp((char*)yytext,"!=") == 0)
|
||||
yylval.str = pstrdup("<>"); /* compatability */
|
||||
|
|
@ -124,49 +206,6 @@ other .
|
|||
CheckFloat8Val(yylval.dval);
|
||||
return (FCONST);
|
||||
}
|
||||
{quote} {
|
||||
char literal[MAX_PARSE_BUFFER];
|
||||
int i = 0;
|
||||
int c = 0;
|
||||
/* quote_seen can be either \ or ' because
|
||||
we handle both cases of \' and '' for
|
||||
quoting quotes*/
|
||||
int quote_seen = 0;
|
||||
|
||||
while (i < MAX_PARSE_BUFFER - 1) {
|
||||
c = input();
|
||||
if (quote_seen != 0) {
|
||||
if (quote_seen == '\'' &&
|
||||
c != '\'') {
|
||||
/* a non-quote follows a single quote */
|
||||
/* so we've hit the end of the literal */
|
||||
if (c != '\0' && c != EOF)
|
||||
unput(c); /* put back the extra char we read*/
|
||||
i = i - 1;
|
||||
break; /* break out of the while loop */
|
||||
}
|
||||
/* if we reach here, we're still in */
|
||||
/* the string literal */
|
||||
literal[i++] = c;
|
||||
quote_seen = 0;
|
||||
continue;
|
||||
}
|
||||
if (c == '\0' || c == EOF) {
|
||||
elog(WARN,"unterminated quoted string literal");
|
||||
/* not reached */
|
||||
}
|
||||
literal[i++] = c;
|
||||
if (c == '\'' || c == '\\')
|
||||
quote_seen = c;
|
||||
}
|
||||
if ( i == MAX_PARSE_BUFFER - 1) {
|
||||
elog (WARN, "unterminated quote string. parse buffer of %d chars exceeded", MAX_PARSE_BUFFER);
|
||||
/* not reached */
|
||||
}
|
||||
literal[i] = '\0';
|
||||
yylval.str = pstrdup(scanstr(literal));
|
||||
return (SCONST);
|
||||
}
|
||||
{identifier} {
|
||||
int i;
|
||||
ScanKeyword *keyword;
|
||||
|
|
@ -177,19 +216,25 @@ other .
|
|||
|
||||
keyword = ScanKeywordLookup((char*)yytext);
|
||||
if (keyword != NULL) {
|
||||
if ( keyword->value == DEFAULT )
|
||||
if ( keyword->value == DEFAULT ) {
|
||||
DefaultStartPosition = CurScanPosition () + yyleng + 1;
|
||||
else if ( keyword->value == CHECK )
|
||||
printf( "default offset is %d\n", DefaultStartPosition);
|
||||
|
||||
} else if ( keyword->value == CHECK ) {
|
||||
CheckStartPosition = CurScanPosition () + yyleng + 1;
|
||||
printf( "check offset is %d\n", CheckStartPosition);
|
||||
|
||||
};
|
||||
|
||||
return (keyword->value);
|
||||
} else {
|
||||
yylval.str = pstrdup((char*)yytext);
|
||||
return (IDENT);
|
||||
}
|
||||
}
|
||||
{space} { /* ignore */ }
|
||||
{space} { /* ignore */ }
|
||||
|
||||
{other} { return (yytext[0]); }
|
||||
{other} { return (yytext[0]); }
|
||||
|
||||
%%
|
||||
|
||||
|
|
@ -282,7 +327,12 @@ myinput(char* buf, int max)
|
|||
int
|
||||
CurScanPosition(void)
|
||||
{
|
||||
printf( "current position is %d\n", yy_c_buf_p - yy_current_buffer->yy_ch_buf - yyleng);
|
||||
|
||||
return (parseCh - parseString - yyleng - 1);
|
||||
#if FALSE
|
||||
return (yy_c_buf_p - yy_current_buffer->yy_ch_buf - yyleng);
|
||||
#endif
|
||||
}
|
||||
|
||||
#endif /* FLEX_SCANNER */
|
||||
|
|
|
|||
Loading…
Reference in New Issue