| Commit message (Collapse) | Author | Age |
... | |
| | | |
|
| | | |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
- echo "" does not hang in powershell
- cmd.exe's echo command does not hang.
- job tests default to powershell (WHY?)
- wait 5 seconds for powershell to create an empty file
- powershell is slow
- cannot reliably validate the id returned by jobstart via jobpid, jobstop
- if using cmd.exe, waiting for a second should be enough
- remaining job tests are unreliable in Windows because any build can pass/fail
for same conditions without changes, especially if the error is in stderr
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Problem: Completion items cannot be annotated.
Solution: Add a "user_data" entry to the completion item. (Ben Jackson,
coses vim/vim#2608, closes vim/vim#2508)
https://github.com/vim/vim/commit/9b56a57cdae31f7a2c85d440392bf63d3253a158
|
| | |
| | |
| | |
| | |
| | |
| | | |
It was added in the parent commit, but ended up not being used. And
I can't think of a case where it will be used: instead we would probably
want to generalize expect_msg_seq() if necessary.
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
job_spec.lua on AppVeyor (Windows) often fails like this:
FAILED ] C:/projects/neovim/test/functional\core\job_spec.lua @ 72: jobs changes to given `cwd` directory
C:/projects/neovim/test/functional\core\job_spec.lua:81: Expected objects to be the same.
Passed in:
(table) {
[1] = 'notification'
[2] = 'stdout'
*[3] = {
[1] = 0
*[2] = {
[1] = 'C:\projects\neovim\Xtest-tmpdir\nvimmSjq1S\0' } } }
Expected:
(table) {
[1] = 'notification'
[2] = 'stdout'
*[3] = {
[1] = 0
*[2] = {
[1] = 'C:\projects\neovim\Xtest-tmpdir\nvimmSjq1S\0'
*[2] = '' } } }
stack traceback:
Message chunking is non-deterministic, so we need to try different
variants.
|
| |/ |
|
| |\
| | |
| | | |
ex_getln: clear cmdline_block after it's freed
|
| | | |
|
| |/ |
|
| |
| |
| |
| |
| | |
Developer can use :checkhealth to get more details, don't need to blast
the details in the skip-message every time.
|
| | |
|
| |
| |
| |
| |
| |
| |
| | |
Test_edit_08() depends on special-case handling in has_compl_option()
and redrawing() which are in Vim but are not wanted in Nvim. Using a Lua
test instead of depending on workarounds in the core to make the VimL
test work.
|
| |
| |
| |
| |
| |
| |
| |
| | |
Problem: Using CTRL-G with 'incsearch' and ? goes in the wrong direction.
(Ramel Eshed)
Solution: Adjust search_start. (Christian Brabandt)
https://github.com/vim/vim/commit/da5116da4586fc714434411d2cccb066caa3723e
|
| |\ |
|
| | |\
| | | |
| | | |
| | | | |
closes #7937
|
| | | | |
|
| |/ /
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Make HlAttr contain highlighting state for both color modes (cterm and rgb).
This allows us to implement termguicolors completely in the TUI.
Simplify some logic duplicated between ui.c and screen.c. Also avoid
some superfluous highlighting reset events.
|
| | | |
|
| | | |
|
| | |
| | |
| | |
| | | |
they test the same thing. Filtering is tested elsewhere.
|
| | | |
|
| | |
| | |
| | |
| | | |
ref #6289
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
closes #5442
closes #4142
ref #6618
ref #4376
ref #7844
ref #2958
ref #4338
|
| |/
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The old behavior is probably not justified, for the usual reason:
terminal buffers may have interactive processes, so cursor placement is
arbitrary, therefore tracking it in the jumplist is useless (or worse).
N.B.: per the docstring for `checkpcmark()` it looks like we were
calling `checkpcmark()` and `setpcmark()` in the wrong order.
closes #3723
|
| | |
|
| |
| |
| |
| | |
And similarly nvim_command_output test
|
| |
| |
| |
| | |
Also update existing tests for new (vim-compatible) newline behavior
|
| | |
|
| |
| |
| | |
It appears to be different on lua and luajit.
|
| |
| |
| | |
May be needed for unit tests as well though.
|
| |
| |
| |
| |
| |
| | |
Lua (not LuaJIT) complains about the "^[[" strings inside the expect,
since it sees them as nested quotes. Change the quoting to [=[ ]=] to
avoid the issue.
|
| |\ |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Problem: CTRL-G/CTRL-T don't work with incsearch and empty pattern.
Solution: Use the last search pattern. (Christian Brabandt, closes vim/vim#2292)
https://github.com/vim/vim/commit/d0480097177369a6ed91d47aba189ae647afcd68
|
| | | |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Problem: Incremental search only shows one match.
Solution: When 'incsearch' and and 'hlsearch' are both set highlight all
matches. (haya14busa, closes vim/vim#2198)
https://github.com/vim/vim/commit/2e51d9a0972080b087d566608472928d5b7b35d7
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Use unique filenames to avoid test conflicts.
Use read_file() instead of io.popen(), to ensures the file is closed.
Use helpers.rmdir(), it is far more robust than lfs.
closes #7911
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Test failure:
test/functional/eval/system_spec.lua: "works with an empty string"
E5677: Error writing input to shell-command: EPIPE
ref https://github.com/neovim/neovim/pull/6558#issuecomment-361061035
ref #6554
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
vim-patch:8.0.0358: invalid memory access in C-indent code
Problem: Invalid memory access in C-indent code.
Solution: Don't go over end of empty line. (Dominique Pelle, closes vim/vim#1492)
https://github.com/vim/vim/commit/60629d642541a089c322e65963c0a77e5f77eb79
vim-patch:8.0.0359: 'number' and 'relativenumber' are not properly tested
Problem: 'number' and 'relativenumber' are not properly tested.
Solution: Add tests, change old style to new style tests. (Ozaki Kiichi,
closes vim/vim#1447)
https://github.com/vim/vim/commit/dc9a081712ec8c140e6d4909e9f6b03a629d32d3
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Per CMAKE docs, CMAKE_HOST_SYSTEM_VERSION is the result of `uname -r`:
https://cmake.org/cmake/help/v3.4/variable/CMAKE_HOST_SYSTEM_VERSION.html?highlight=uname
A numeric version string for the system. On systems that support
uname, this variable is set to the output of uname -r. On other
systems this is set to major-minor version numbers.
On Windows it is something like "6.1", so it won't match ".*-Microsoft".
Closes #7329
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Problem: Terminal width is set to 80 in test3.
Solution: Instead of setting 'columns' set 'wrapmargin' depending on
'columns.
https://github.com/vim/vim/commit/38a3d6c9601b637a28f399059263300e9f65eba4
|
| |\ \ |
|
| | | | |
|
| | | | |
|
| | | | |
|
| | | | |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
This reverts commit f7fe3012204169f22412194a78f196ffc72bb8c3.
Fails on QuickBuild because it uses a non-local path.
Need a environment-agnostic solution
|
| | | | |
|
| | | | |
|
| | | | |
|