Problem: Using the system default encoding makes tests unpredictable.
Solution: Always use utf-8 or latin1 in the new style tests. Remove setting
encoding and scriptencoding where it is not needed.
ac105ed3c4
Attempting to serialize a NULL string through msgpack results in
msgpack_sbuffer_write attempting to memcpy from a NULL pointer, which is
undefined behavior.
The existing code would cause utf8len_tab to be declared as non-extern
when main.cpp included globals.h as well as in mbyte.c. This causes the
following warning
Linking C executable ../../bin/nvim
/usr/bin/ld: Warning: size of symbol `utf8len_tab' changed from 256 in CMakeFiles/nvim.dir/main.c.o to 320 in CMakeFiles/nvim.dir/mbyte.c.o
Moving the definition to globals.h and using INIT() ensures the array is
only defined in main.cpp and other places globals.h is included see an
extern declaration.
This fixes failures with TSAN builds like
FATAL: ThreadSanitizer can not mmap the shadow memory (something is mapped at 0x55deea465000 < 0x7cf000000000)
FATAL: Make sure to compile with -fPIE and to link with -pie.
Problem: Command line editing is not tested much.
Solution: Add tests for expanding the file name and 'wildmenu'.
ae3150ec8d
The test_cmdline.vim tests were ported in 23f591dba0 and thus
should have marked 1923 as applied. The test_cmdline.vim invocation
has been moved from test_alot.vim to src/nvim/testdir/Makefile to
better accord with the upstream code.
Problem: When switching to a new buffer and an autocommand applies syntax
highlighting an ml_get error may occur.
Solution: Check "syn_buf" against the buffer in the window. (Alexander von
Buddenbrock, closesvim/vim#676)
b681be175b
Problem: When using :stopinsert in a silent mapping the "INSERT" message
isn't cleared. (Coacher)
Solution: Always clear the message. (Christian Brabandt, closesvim/vim#718)
fd773e9e88
Problem: Overwriting pointer argument.
Solution: Assign to what it points to. (Dominique Pelle)
76ae22fef3
The typos corrected in the original vim patch are no long present
in the neovim code base and the pointer assignment was done correctly
in the porting of patch 1913 (where the changes were introduced).
Problem: When setting 'filetype' there is no check for a valid name.
Solution: Only allow valid characters in 'filetype', 'syntax' and 'keymap'.
d0b5138ba4
https://github.com/mpeterv/luacheck/pull/81#issuecomment-261099606
> If you really want to use bleeding-edge version you should get the
> rockspec from master branch, not a fixed commit ...
> The correct way to install from a specific commit is cloning that
> commit and running "luarocks make" from project directory. The reason
> is that running "install" or "build" on an scm rockspec fetches
> sources from master but uses build description from the rockspec
> itself, which may be outdated.
* Fix translation of the term 'tag'
Prefer native word instead of transliteration
* Update uk.po automatically
Run make update-po-uk in src/nvim/po
* Update fuzzy translations in uk.po
* Translate missing msgstr
* Fix layout of the front page
* Apply cleanup.vim
Additionally, remove irrelevant and confusing comments
* Copy uk.po into uk.cp1251.po
Problem: Using freed memory after setqflist() and ":caddbuffer". (Dominique)
Solution: Set qf_ptr when adding the first item to the quickfix list.
8b20179c65
In order to not conflict with new error codes that Vim adds, all Neovim
error codes should be above 5000. The three existing sub-5000 error
codes (E926, E951, and E952) are now E50003, E5004, and E5005
respectively.
E953 was removed in 6167ce6df2, so just
remove it from the help.
Problem: Crash when an autocommand changes a quickfix list. (Dominique)
Solution: Check wether an entry is still valid. (Yegappan Lakshmanan,
Hirohito Higashi)
ffec3c5349
Problem: Order of matches for 'omnifunc' is messed up. (Danny Su)
Solution: Do not set compl_curr_match when called from complete_check().
(closesvim/vim#1168)
472e85970e
grep 2.26 changed its behavior when redirecting its output to /dev/null
such that it exits as soon as one match is found. This causes sed to
get a SIGPIPE which, due to "set -o pipefail", falsely implies that the
patch is not applied.
Removing "set -o pipefail" preserves the good exit status from grep.