- use CACHE_NVIM_DEPS_DIR
- do not cache pip
This is handled through http caches in general/better, and it is not
used much anyway.
- do not cache DEPS_DOWNLOAD_DIR
Built deps are cached, downloads are not needed then.
- display ccache stats before clearing
- do not cache ccache stats
- improve output of `du` (do not list pages of output for "/home/travis/.cache/go-build")
It should be done for timeouts only (-1, not != -1), but the job is
stopped via `:bwipeout!` already also in that case.
Fixes E900 with timing out tests:
"test_xxx.tlog" "test_xxx.tlog" [New] 183L, 13375C written
Error detected while processing function Main:
line 37:
E900: Invalid channel id================================================================================
* scripts/vim-patch.sh: fall back to "origin" for remote
Without this, it would fail e.g. with a locally cloned repo of Neovim.
* scripts/vim-patch.sh: assign_commit_details: handle tags [ci skip]
- Handle "v" prefix from Vim tags.
- Exit in case of error therein already.
* -l/-L: pass through git-log options [ci skip]
This allows for only listing missing patches for a given Vim file:
> scripts/vim-patch.sh -L src/edit.c
Problem: unreliable types for complete function arguments
Solution: fix argument type for functions w/ unreliable type conversion(Ozaki Kiichi)
vim/vim#2993
Problem: Too many curly braces.
Solution: Remove curly braces where they are not needed. (Hirohito Higashi,
closesvim/vim#3982)
abab0b0fdd
Neovim code style requires the opposite.
Add curly braces to minimize lint errors when applying Vim patches.
Problem: When user tries to exit with CTRL-C message is confusing.
Solution: Only mention ":qa!" when there is a changed buffer. (closesvim/vim#4163)
a84a3dd663
vim-patch:8.1.1052: test for CTRL-C message sometimes fails
Problem: test for CTRL-C message sometimes fails
Solution: Make sure there are no changed buffers.
553e5a5c56
vim-patch:8.1.1053: warning for missing return statement
Problem: Warning for missing return statement. (Dominique Pelle)
Solution: Add return statement.
d6c3f1fa2b
vim-patch:8.0.1753: various warnings from a static analyser
Problem: Various warnings from a static analyser
Solution: Remove unneeded conditions.
1c17ffa461
- Check shada_pack_xx() result, abort on failure.
- Use xcalloc() to zero-initialize `wms`, fixes below ASAN failure.
ASAN failure (running vim_spec.lua in a loop):
../src/nvim/shada.c:1773:13: runtime error: load of value 224, which is not a valid value for type 'bool'
SUMMARY: UndefinedBehaviorSanitizer: undefined-behavior ../src/nvim/shada.c:1773:13
Adding an explicit check just before the ASAN failed line:
case kSDItemRegister: {
if (*((int *)(void *)&entry.data.reg.is_unnamed) != 1
&& *((int *)(void *)&entry.data.reg.is_unnamed) != 0) {
abort();
(gdb) p entry.data.reg
+p entry.data.reg
$5 = {name = 49 '1', type = kMTLineWise, contents = 0x60200000c250, is_unnamed = true, contents_size = 1, width = 0, additional_data = 0x0}
(gdb) p *((uint8_t *)(void *)&entry.data.reg.is_unnamed)
+p *((uint8_t *)(void *)&entry.data.reg.is_unnamed)
$7 = 2049
Problem: Cannot highlight the "~" lines at the end of a window differently.
Solution: Add the EndOfBuffer highlighting. (Marco Hinz, James McCoy)
58b853460a
Includes minor doc parts from 7.4.1304 for runtime/doc/eval.txt.
Having llvm-symbolizer in the $PATH is enough.
- check_logs: remove log after displaying it
Otherwise it would be displayed/symbolized again and again.
E.g. in https://api.travis-ci.org/v3/job/564477704/log.txt.
* before changed line status variable is set to RA_FAIL in 2 places
line 3816 and 3826, both lead to short-circuit of the loop, making
comparing it to RA_FAIL useless
close#10624
Since timed_out is pointer to int condition timed_out != NULL && timed_out makes no sense as both are effectively the same thing. Most likely the first check was meant to guard the dereference of the second.
Wrong import of Vim patch (8.0.0643).
Daniel Hahler