| Commit message (Collapse) | Author | Age |
|\
| |
| |
| | |
7.4.2313, 7.4.2314'.
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Also make setpos("'A", [999, 1, 1, 0]) fail, i.e. return -1 (assuming there is no buffer 999).
Fixes #5713
Background:
`:help setpos()` mentions an argument `"bufnum"` that determines the buffer a mark should be put in.
This argument is respected for uppercase marks, but not for lowercase marks.
This is reasonable (though I personally would like `setpos()` to be able to set marks in other buffers), but the help doesn't mention this anywhere.
It's also strange that attempting to change buffers with `setpos('.', [bufnr('#'), 1, 1, 0])` alerts the user that having a different buffer is an error, while attempting to set a mark with `setpos("'d", [bufnr('#'), 1, 1, 0])` doesn't tell the user that the `"bufnum"` argument is an error.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
While a job callback is active, it may be invoked again. Since the
data handled by the first invocation of the callback hasn't been marked
as consumed, the subsequent invocation will see the same data.
Reported-by: Daniel Hahler
Patch-by: oni-link
Closes #5889
|
|\ \
| | |
| | | |
open_buffer(): Raise `BufEnter` for directories.
|
| | | |
|
| | | |
|
|/ /
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This fixes a use-after-free noticed by ASAN which would occur when a
dictwatcher was still active on a dictionary when the dictionary was
freed.
fun! MakeWatch()
let d = {'foo': 'bar'}
call dictwatcheradd(d, 'foo', function('...'))
endfun
Patch-by: oni-link
Closes #5930
|
| |
| |
| |
| | |
Closes #5888
|
| |
| |
| |
| | |
Closes #5912
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| | |
Do _not_ set v:shell_error on parameter validation error.
system([...]) does not invoke a shell, so this change is somewhat
questionable. But `:help v:shell_error` is sufficiently vague to allow
-1 in this case.
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| | |
Problem: evalcmd() has a confusing name.
Solution: Rename to execute(). Make silent optional. Support a list of
commands.
https://github.com/vim/vim/commit/79815f1ec77406f2f21a618c053e5793b597db7a
|
| |
| |
| |
| |
| | |
One of the tests remains disabled until we figure out the expected
behaviour of mkdir() on Windows when `prot` is passed.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
After #4964 environment variables in the XDG "fallback" table are no
longer expanded.
Fallback to correctly expanded $LOCALAPPDATA, $TEMP. If that fails
(unlikely), fallback to hard-coded paths (e.g. ~/AppData/Local).
Closes #5255
|
|\ \
| | |
| | | |
shada: Save current cursor position before saving jumps
|
| | | |
|
| | | |
|
| | | |
|
| | | |
|
| | | |
|
| | | |
|
| |/
|/| |
|
|\ \
| | |
| | | |
Refactor eval/typval_encode.h
|
| | | |
|
| | | |
|
| | |
| | |
| | |
| | |
| | | |
Also fixed dumping of partials by encode_vim_to_object and added code which is
able to work with partials and dictionaries to test/unit/eval/helpers.lua
(mostly copied from #5119, except for partials handling).
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Occurs when trying to dump a partial with attached self dictionary which
references that partial. “Infinite” loop should normally result in Neovim killed
by OOM killer.
Also moved the place when partials are unreferenced by clear_tv: from
…FUNC_START to …FUNC_END.
|
| | |
| | |
| | |
| | |
| | |
| | | |
- Clarify documentation.
- Return `FAIL` and truncate if `fname` is too long.
- Add tests.
|
|\ \ \
| | | |
| | | | |
Fix plugin which opens ShaDa files
|
| | | |
| | | |
| | | | |
Fixes #5482
|
| | |/
| |/| |
|
|\ \ \
| |/ /
|/| | |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Let build systems define TEST_SKIP_FRAGILE to skip tests that are known to be
resource-intensive (unreliable on slow systems).
References https://github.com/neovim/neovim/pull/5488#issuecomment-265622113
|
| |/
| |
| |
| |
| |
| |
| | |
- Improve test reliability by only checking for a line with the string
we are interested in ("Interrupt").
- Try to avoid OOM by loading an existing big file instead of looping to
create one.
|
|\ \ |
|
| | |
| | |
| | |
| | |
| | |
| | | |
On Windows, if the nvim process has a directory open the lua process
cannot remove it. After failing once, it's safe to force `nvim` to the
top-level directory. Then try again.
|
| |/
| |
| |
| |
| |
| |
| | |
- Eliminate global test_autochdir.
- Eliminate VimL function test_autochdir()
- Use a lua test instead. Fails correctly after reverting
0c4347997954 / vim-patch:7.4.2015.
|
| | |
|
| |
| |
| |
| |
| |
| | |
Keeps arguments separated and not joined as a single string as long as possible.
Abstracts away additional arguments so that Gcc:preprocess should work for
compilers with different conventions should they be supported.
|
| | |
|
| |
| |
| |
| | |
Previous commit made preprocess.lua know how its output will be used. This moves
state to cimport, making only it know which is cleaner.
|
|/
|
|
|
|
|
|
|
|
| |
Works by saving all preprocessor defines and reusing them on each run. This also
saves NVIM_HEADER_H defines. Saving other defines is needed for defines like
`Map(foo, bar)` which are sometimes used to declare types or functions. Saving
types or function declarations is not needed because they are recorded as luajit
state.
Fixes #5857
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
stdout activity may reach the queue before `exit`; just discard it.
Observed on Travis CI, ubuntu "trusty" beta image:
[ RUN ] ...is/build/neovim/neovim/test/functional/core/job_spec.lua @ 509: jobs running tty-test program jobclose() sends SIGHUP
...is/build/neovim/neovim/test/functional/core/job_spec.lua:511: Expected objects to be the same.
Passed in:
(table) {
[1] = 'notification'
*[2] = 'stdout'
[3] = {
[1] = 0
[2] = {
[1] = '
'
[2] = '' } } }
Expected:
(table) {
[1] = 'notification'
*[2] = 'exit'
[3] = {
[1] = 0
[2] = 42 } }
stack traceback:
...is/build/neovim/neovim/test/functional/core/job_spec.lua:511: in function <...is/build/neovim/neovim/test/functional/core/job_spec.lua:509>
[ FAILED ] ...is/build/neovim/neovim/test/functional/core/job_spec.lua @ 509: jobs running tty-test program jobclose() sends SIGHUP (2.81 ms)
|
|
|
|
| |
Fixes #5455.
|
|
|
|
|
|
|
|
| |
Problem: Arguments of setqflist() are not checked properly.
Solution: Add better checks, add a test. (Nikolai Pavlov, Hirohito Higashi,
closes vim/vim#661)
https://github.com/vim/vim/commit/d106e5ba7f10f0d2a14eaefe5d78405044416cb9
|
|
|
|
|
| |
Massaging the upstream patches for this test into the lua tests are too
cumbersome and slow down patching.
|
| |
|
| |
|