Problem: With 'autochdir' win_execute() can corrupt the buffer name,
causing :write to use wrong path.
Solution: Save and restore b_fname when 'autochdir' is active
(Ingo Karkat).
This is caused by a bad interaction of the 'autochdir' behavior,
overriding of the current directory via :lchdir, and the temporary
window switching done by win_execute(), manifesting when e.g. a custom
completion inspects other buffers:
1. In the initial state after the :lcd .. we have curbuf->b_fname =
"Xsubdir/file".
2. do_autochdir() is invoked, temporarily undoing the :lcd .., changing
back into the Xsubdir/ subdirectory.
3. win_execute() switches windows, triggering win_enter_ext() →
win_fix_current_dir() → shorten_fnames(TRUE)
4. shorten_fnames() processes *all* buffers
5. shorten_buf_fname() makes the filename relative to the current
(wrong) directory; b_fname becomes "file" instead of "Xsubdir/file"
6. Directory restoration correctly restores working directory via
mch_chdir() (skipping a second do_autochdir() invocation because
apply_acd is FALSE), but b_fname remains corrupted, with the
"Xsubdir/" part missing.
7. expand("%:p") (and commands like :write) continue to use the
corrupted filename, resolving to a wrong path that's missing the
"Xsubdir/" part.
To fix the problem the short filename is saved if its in effect (i.e.
pointed to by curbuf->b_fname) and 'autochdir' happened. It's then
restored in case of a local cwd override. The conditions limit this
workaround to when 'autochdir' is active *and* overridden by a :lchdir.
closes: vim/vim#19343abb4d74033
Co-authored-by: Ingo Karkat <swdev@ingo-karkat.de>
Problem: Some tests are not valid on OpenBSD.
Solution: Add CheckNotOpenBSD, use it to skip certain tests
(Kevin Goodsell).
Test_readdirex_sort performs locale-dependent sorting. OpenBSD has
minimal locale support.
Test_stdin_no_newline hangs on OpenBSD and FreeBSD. I don't know exactly
why, but it may be due to bash not exiting at the end of the test. This
is skipped in the FreeBSD CI runs because bash is not installed.
Test_detect_fifo uses /dev/fd/ files (via process substitution) as
FIFOs. On OpenBSD the files in /dev/fd are not FIFOs.
closes: vim/vim#19351a24cb278bd
Co-authored-by: Kevin Goodsell <kevin-opensource@omegacrash.net>
Problem: :tab sbuffer may close old tabpage if BufLeave autocommand
splits window (after 9.1.0143).
Solution: Only close other windows if the buffer will be unloaded
(zeertzjq).
related: neovim/neovim#37749closes: vim/vim#193526da9f757c4
Problem: windows may scroll horizontally upon resize using the old terminal
size, which may be unnecessary and cause the content to be partially out-of-view.
Solution: reset the horizontal scroll after resizing.
Problem:
$NVIM_APPNAME was not respected when searching $XDG_CONFIG_DIRS for
config files. Nvim hardcoded "nvim" when constructing paths like
`$XDG_CONFIG_DIRS/nvim/init.lua`, ignoring the $NVIM_APPNAME environment
variable.
This meant that config files like `$XDG_CONFIG_DIRS/myapp/init.lua` were
not loaded, even though $NVIM_APPNAME was set to "myapp".
Solution:
Use `get_appname()` instead of hardcoded "nvim" for $XDG_CONFIG_DIRS
paths in `do_system_initialization()` and `do_user_initialization()`.
This makes $XDG_CONFIG_DIRS behave consistently with $XDG_CONFIG_HOME,
which already respected $NVIM_APPNAME.
As documented in `runtime/doc/starting.txt` (L1440-L1441):
"In the help wherever `$XDG_CONFIG_…/nvim` is mentioned it is understood
as `$XDG_CONFIG_…/$NVIM_APPNAME`."
See:
43339dee40/runtime/doc/starting.txt (L1440-L1441)
Relates to #37405
Problem:
The E5422 error message was duplicated in two places in main.c, and
hardcoded error strings should be centralized in errors.h for consistency.
Solution:
Add e_conflicting_configs to errors.h and replace both inline error
strings with the centralized constant.
Problem:
`init.lua` files in `$XDG_CONFIG_DIRS` directories were not being sourced during startup, even though the documentation states they should be searched alongside `init.vim`.
See:
e51f5e17e1/runtime/doc/starting.txt (L495-L496)
Solution:
Modify `do_user_initialization()` to search for `init.lua` in each `$XDG_CONFIG_DIRS` directory before falling back to `init.vim`, matching the behavior for `$XDG_CONFIG_HOME`. Also show `E5422` error if both `init.lua` and `init.vim` exist in the same directory.
Fixes#37405
Problem: Exmode is missing cmdline_block events.
Solution: Emit cmdline_block_show/append when executing a cmdline during
exmode. Emit cmdline_block_hide when leaving exmode.
The auto-refresh has a bit of a delay so it can happen that when a user
runs `codelens.run` it operates on an outdated state and either
does nothing, or fails.
This changes the logic for `.run` to always fetch the current lenses
before (optional) prompt and execution.
See discussion in https://github.com/neovim/neovim/pull/37689#discussion_r2764235931
This could potentially be optimized to first check if there's local
state with a version that matches the current buf-version, but in my
testing re-fetching them always was quickly enough that `run` still
feels instant and doing it this way simplifies the logic.
Side effect of the change is that `.run` also works if codelens aren't
enabled - for power users who know what the codelens would show that can
be useful.
Problem: search() is used to check for the message from tar that
indicates leading slashes found in the tar archive, or to
check for the leading slashes themselves. However, if
'nowrapscan' is in effect these searches are limited to the
last line and don't find any results. This causes the warning
message from tar to be seen in the buffer, the "Path Traversal
Attack Detected" message to be omitted, and editing actions
can fail. This can be seen, for example, when editing
src/testdir/samples/evil.tar.
Solution: Use the 'w' flag for search() (Kevin Goodsell)
closes: vim/vim#1933318d844e365
Co-authored-by: Kevin Goodsell <kevin-opensource@omegacrash.net>
Problem: Message ID is not incremented without ext_messages.
Solution: Increment where callstack reaches without ext_messages.
There is no clear place marking the end of internal
messages without ext_messages so IDs are only incremented
with `nvim_echo` calls. Message IDs are currently not exposed
anywhere but the msg_show event for this to matter.
Problem: Unlike `nvim_keymap_set`, `vim.keymap.set` uses the non-negated
`remap` instead of `:set`'s `noremap`, but the documentation for this
got lost sometime before Nvim 0.10.
Solution: Restore the lost documentation and make it more explicit.
Problem: "first" parameter added in 912388f5 for grouping printed lines
in a single event is ineffective for the :global command.
Solution: Only set the message kind for the first printed line during a
:global command.
Problem:
Recombining composing chars in terminal doesn't work at right edge.
Solution:
Check for the case where printing the previous char didn't advance the
cursor. Reset at_phantom when returning to combine_pos.
Problem: Page scrolling in Insert mode beeps (after 9.1.0211).
Solution: Fix incorrect return value of pagescroll(). Also invert the
return value of scroll_with_sms() to be less confusing and
match comments (zeertzjq).
fixes: vim/vim#19326closes: vim/vim#19327a8ce914db1
Problem:
a2b92a5efb is regressive if the
ui_mode_idx set by ui_cursor_shape differs from what ui_flush later sets
it to. (e.g: if mode changes between both calls)
Solution:
Don't overwrite ui_mode_idx to obscure the cursor so that we don't
forget what shape was last set by ui_cursor_shape.
May not be needed, but does fix a potential regression. (and makes it
easier to set ui_mode_idx to something other than cursor_get_mode_idx's
retval in other places, if that's ever wanted)
Problem: UI2 does not implement the msg_show event msg_id parameter.
Solution: Store message IDs currently shown in the cmd/msg buffers.
Set extmarks spanning the message which are used to replace
a still visible message when a new message with the same ID
is received.
Problem:
Users may be unaware that setting `g:clipboard` after providers are
initialized has no effect, and that `has('clipboard')` initializes
providers, as in #13062.
Solution:
Note the restriction and link to workarounds in FAQ for discoverability.
Problem:
There is no indication of when cursor is "behind" an unfocused, floating window.
Solution:
Set underline cursor to indicate when an unfocused floatwin is over the active cursor.
When using vim9-syntax plugin, :VimKeywordPrg does not lookup functions
correctly, as it relies solely on syntax names to find the help topic.
The syntax keyword used for builtin function is vi9FuncNameBuiltin in
vim9-syntax plugin, not vimFuncName expected by :VimKeywordPrg, so the
fallback rules apply, and there is no fallback rule for function calls.
Fix by just checking if the first char after topic is '(', and if so
assume help topic is a function.
closes: vim/vim#19320ac5af8ecd3
Co-authored-by: Mark Woods <mwoods.online.ie@gmail.com>
- Match &option, and &[lg]:option variables.
- Match Ex commands after :bar.
- Fix matching of pre and post context text.
- Style - use '..' for string concatenation.
fixesvim/vim#17567closes: vim/vim#17653c65643cbec
Co-authored-by: Doug Kearns <dougkearns@gmail.com>
Problem:
When building Neovim with Emscripten for WebAssembly, the
`CMAKE_TOOLCHAIN_FILE` is not propagated to dependency builds.
So libuv attempts to include Linux-specific headers like
`sys/epoll.h` which don't exist in the WebAssembly environment.
From #37572:
…/NeovimWeb/external/neovim/.deps/build/src/libuv/src/unix/linux.c:44:10:
fatal error: 'sys/epoll.h' file not found
44 | #include <sys/epoll.h> | ^~~~~~~~~~~~~
Solution:
1. Add initialization for DEPS_CMAKE_ARGS to prevent CMake errors.
2. Implement conditional logic to always pass CMAKE_TOOLCHAIN_FILE to
dependencies.
This allows dependencies like libuv to detect they're building for
WebAssembly and disable incompatible Linux specific features such as
epoll which resolves the sys/epoll.h error.
From the LSP Spec:
> There are two uses cases where it can be beneficial to only compute
> semantic tokens for a visible range:
>
> - for faster rendering of the tokens in the user interface when a user
> opens a file. In this use case, servers should also implement the
> textDocument/semanticTokens/full request as well to allow for flicker
> free scrolling and semantic coloring of a minimap.
> - if computing semantic tokens for a full document is too expensive,
> servers can only provide a range call. In this case, the client might
> not render a minimap correctly or might even decide to not show any
> semantic tokens at all.
This commit unifies the usage of range and full/delta requests as
recommended by the LSP spec and aligns neovim with the way other LSP
clients use these request types for semantic tokens.
When a server supports range requests, neovim will simultaneously send a
range request and a full/delta request when first opening a file, and
will continue to issue range requests until a full response is
processed. At that point, range requests cease and full (or delta)
requests are used going forward. The range request should allow servers
to return a result faster for quicker highlighting of the file while it
works on the potentially more expensive full result. If a server decides
the full result is too expensive, it can just error out that request,
and neovim will continue to use range requests.
This commit also fixes and cleans up some other things:
- gen_lsp: registrationMethod or registrationOptions imply dynamic
registration support
- move autocmd creation/deletion to on_attach/on_detach
- debounce requests due to server refresh notifications
- fix off by one issue in tokens_to_ranges() iteration
