Problem: win_linetabsize() includes wrap overhead from 'linebreak'
based on current window width, but the result sizes the window,
causing a feedback loop.
Solution: Temporarily set w_view_width to Columns before measuring.
Problem: _get_and_set_name edits the name for the whole group,
thus only one client per group gets the didOpen message.
Solution: move the logic to _changetracking and loop over every
client per group.
Problem:
Due to optimizations c936ae0f36, nvim prints literal spaces instead of using
`erase_chars` in widths <= 5 even if the terminal advertises `erase_chars`
support (perhaps a small-output size heuristic). However, this is not
semantically neutral: in some terminals, erased cells and printed spaces are
copied differently.
I ended up with two useful groups of results.
First, I tested raw terminal behavior without nvim involved:
| Terminal | Raw plain text | Raw `erase_chars` | Raw literal spaces |
| --- | --- | --- | --- |
| xterm | clean | trailing spaces copied | trailing spaces copied |
| xfce4-terminal | clean | clean | trailing spaces copied |
Second, I tested nvim itself:
| Terminal | no patch | with this patch |
| --- | --- | --- |
| xfce4-terminal | trailing spaces reproduced | clean |
| xterm | trailing spaces reproduced | trailing spaces reproduced |
| Alacritty | clean | clean |
| Ghostty | clean | clean |
| WezTerm | clean | clean |
Nvim often prints spaces instead of sending `erase_chars`, which this patch
changes for short clears when the terminal advertises it. This fixes
xfce4-terminal because raw `erase_chars` are already cleaned up by the terminal,
while spaces aren't. ***Notably, xterm is different***: even when `erase_chars`
is sent directly (NO NVIM INVOLVED) xterm *still* copies those cleared blank
trailing cells (and this is documented). So for xterm, which is the only
remaining problematic fix, I'm quite sure there's nothing we ought to do on the
Nvim side.
Solution:
Drop the `width >= 5` condition.
Problem:
If `'keywordprg'` begins with `:`, `3K` turns the count into an Ex
range. Commands that don't support that then fail. Vim passes the count
as the first arg (see #19436, vim/vim#10745).
Solution:
Pass `[count]` as the first arg for `'keywordprg'`.
Problem: hlgroup2dict passes &ns_id to ns_get_hl twice. The first call
(link=true) sets *ns_hl = 0 when link_global is set, so the second call
and the sg_cleared guard both see ns_id == 0 and bail out. The group is
silently dropped from the result.
Solution: use a temporary copy of ns_id for each ns_get_hl call so the
original value is preserved.
Problem:
Neovim currently fails to build on GNU Hurd. Because Hurd relies on
glibc, `<sys/param.h>` defines the `BSD` macro for 4.4BSD compatibility.
The preprocessor incorrectly routes Hurd into the BSD code paths, which
fatally fail during compilation because Hurd lacks `<sys/sysctl.h>` and
the `sysctl()` function.
Solution:
Update the preprocessor guards in `os/proc.c` to explicitly exclude
`__gnu_hurd__` from the BSD-specific `sysctl` blocks. Instead, group GNU
Hurd with the `__linux__` paths, as both systems rely on standard POSIX
interfaces and `/proc` parsing (which Hurd fully supports via its
`procfs` translator).
Testing:
The test suite does not fully pass yet natively on GNU Hurd
(specifically tests involving PTY closures and SIGHUP/SIGTERM trapping,
like `autocmd TermClose kills PTY job`). This is due to underlying
differences in Hurd's Mach RPC architecture and the `term` translator.
This patch does not attempt to fix those test executions, but simply
unblocks the core compiler as a necessary first step.
Problem:
The ui compositor does not use grid_scroll events when a grid other than
the built in msg_grid exists above the scrolled grid, regardless of
whether it actually intersects the scrolled grid anywhere in the
scrolled region.
When another layer exists at a higher zindex, the ui compositor falls
back to composing every line of the scrolled grid. This is especially
evident when using ui2 which creates a floating window that replaces the
built-in msg_grid. Scrolling around with ui2 enabled has poor
performance because the entire grid is recomposed on every scroll
instead of using grid_scroll.
Solution:
Instead of just checking whether another grid exists at a higher zindex
than curgrid, ensure there's a grid above the curgrid that is positioned
over the particular rectangle of interest, which could be a subrectangle
of the whole grid.
Grids above the curgrid that don't intersect there no longer count as
covering the curgrid, and the compositor can continue to use grid_scroll
scroll events. The floating window created for the command line/messages
with ui2 never intersects with the scrolled region of the main window
grid, which allows performant scrolling with it enabled.
Problem: Assuming modifyOtherKeys for rhs of mapping.
Solution: Ignore seenModifyOtherKeys for mapped characters. (closesvim/vim#6200)
46cd43bda1
----
"getchar.c" changes depend on patch 8.1.2145.
Can't port it due to tests.
"test_gui.vim" doesn't depend on GUI for all tests.
----
Co-authored-by: Bram Moolenaar <Bram@vim.org>
Problem:
On Git versions 2.13..2.26 there is a bug that prevents using
`stash --message`.
Solution:
Use the full `stash push --message` form to avoid that bug.
Problem: hover/signature callback lacked consistency checks, so slow LSP servers
could open a float after the cursor had already moved away.
Solution: guard the callback with buf validity, buf version, and cursor
position checks before opening the float. Also fix table capacity calculation.
Problem:
When a new textDocument/codeLens response arrives with unresolved lenses,
on_win clears the existing codelens row before codeLens/resolve
completes. This causes the displayed codelens text to flicker while
typing.
Solution:
Keep the current virtual lines if any of the refreshed lenses are still
unresolved. Clear the old virtual lines only when the line no longer has
lenses or all its lenses are resolved.
A trade-off is that the user may temporarily see outdated codelenses.
However, that's preferable to spamming updates on every refresh.
AI-assisted: Codex
Problem: Breaking a link with update=true loses colors inherited from
the linked group.
Solution: Copy color indices from the linked group so inherited colors
remain visible in :hi output.
vim-patch:9.2.0313: Callback channel not registered in GUI
vim-patch:9.2.0319: popup: rendering issues with partially transparent popups
vim-patch:9.2.0322: tests: test_popupwin fails
vim-patch:3e194b106 runtime(vimball): detect more path traversal attacks
vim-patch:9.2.0335: json_encode() uses recursive algorithm
vim-patch:9.2.0309: Missing out-of-memory check to may_get_cmd_block()
vim-patch:9.2.0310: unnecessary work in vim_strchr() and find_term_bykeys()
vim-patch:8.2.2824: MS-Windows: build failure with MSVC
vim-patch:9.1.1692: global_functions are not constant
Problem: Missing help for a function goes unnoticed.
Solution: Add a test. (Gary Johnson)
6b0e528368
----
Test relies on parsing runtime/doc/*.txt and src/evalfunc.c .
Error-prone because Vim and Nvim diverged on both documentation
and source code in spite of ported Vim patches.
Importing src/nvim/eval.lua is an alternative
but it uses associatve index such that it's unsorted.
Treat it as N/A because I can't rewrite any of its tests
under minimal effort.
----
Co-authored-by: Bram Moolenaar <Bram@vim.org>
Problem: spell_read_aff() uses sprintf() into a fixed-size stack buffer
without bounds checking. store_aff_word() uses STRCAT() to
append attacker-controlled strings into newword[MAXWLEN] without
checking remaining space. Both are reachable via :mkspell with
crafted .aff/.dic files (xinyi234)
Solution: Replace sprintf() with vim_snprintf() in spell_read_aff().
Replace STRCAT() with STRNCAT() with explicit remaining-space
calculation in store_aff_word().
closes: vim/vim#1994407faa961a0
Co-authored-by: Christian Brabandt <cb@256bit.org>
Problem: Some patterns in tar and zip plugin tests not strict enough.
Solution: Use assert_equal() for lines that should match exactly. Match
a literal dot properly (zeertzjq).
closes: vim/vim#199462fbc69c9ad
Problem: If the lockfile points to the revision that is not on disk, the
`:checkhealth vim.pack` suggests to run `vim.pack.update()`. Although
usually it should resolve the problem, it is not always the case: like
if the state on disk is already the latest available.
Solution: Add an extra suggestion with a more drastic measure by
manually removing `rev` field from the lockfile for it to be repaired
after the `:restart`.
neovim always uses encoding=utf8. Thus a lot of utf8 tests are just a
lot of tests. Meanwhile test_regexp_latin.vim is a bit of a lone child,
it can just run as a separate test without a shell layer (running it in
the same process as its utf8 sibling is not allowed
Problem: runtime(tar): but with dotted path
Solution: Do not strip everything after the first dot
(Aaron Burrow)
tar#Extract was getting the extensionless basename by
stripping away everything starting with the leftmost
dot. So if a directory had a dot or the file had an
'extra' dot then the code did the wrong thing. For
example, if it was given:
/tmp/foo.bar/baz.tar.gz
Then it would treat /tmp/foo as the extensionless
basename, but it actually should have grabbed:
/tmp/foo.bar/baz
This patch fixes the issue by instead looking at the
rightmost dot(s).
This bug was discovered by ChatGPT 5.4. I wrote the
patch and tested vim.
closes: vim/vim#199304a1bcc67b4
Co-authored-by: Aaron Burrow <burrows@fastmail.com>
Problem: patch 9.2.0325: runtime(tar): bug in zstd handling
Solution: use correct --zstd argument, separated from other arguments,
rework testing framework (Aaron Burrow).
The tar.vim plugin allows vim to read and manipulate zstd archives,
but it had a bug that caused extraction attempts to fail.
Specifically, if the archive has a .tar.zst or .tzst extension, then
the code was generating invalid extraction commands that looked like
this:
tar --zstdpxf foo.tar.zst foo
When they should be like this:
tar --zstd -pxf foo.tar.zst foo
This patch changes the flag manipulation logic so that --zstd isn't
glued to pxf.
The labor for this change was divided between ChatGPT 5.4 and me.
ChatGPT 5.4 identified the issue (from a code scan?), and I wrote
the patch and tested vim.
related: vim/vim#1993000285c035a
Note: tests need the next patch to pass in Nvim.
Co-authored-by: Aaron Burrow <burrows@fastmail.com>
The regex for status line highlighting was too broad, `jjComment` lines
containing e.g. the letter 'A' followed by a space anywhere in the line
were highlighted.
related: vim/vim#19879959817472d
Co-authored-by: Joël Stemmer <jstemmer@google.com>
Problem: tests: test_indent.vim leaves swapfiles behind
Solution: Close open buffer using :bw! instead of :close!
158947e294
Co-authored-by: Christian Brabandt <cb@256bit.org>
Problem: 0x9b byte not unescaped in <Cmd> mapping (BenYip).
Solution: Translate K_CSI to CSI like what is done in vgetc().
(zeertzjq).
fixes: vim/vim#19936closes: vim/vim#199373e2012914e
Problem:
`vim.lsp.buf.definition`/`vim.lsp.buf.declaration` use the same underlying code
via `get_locations`, whereas `vim.lsp.buf.reference` does not. This is because
`buf.reference` does not perform a jump when there is only one item.
Solution:
In #38510, I simplified the jump logic using `:cfirst`, so they can now share
code more easily. Additionally, this PR enables `buf.definition` to display the
corresponding qflist name.
Problem:
`:help dev-name-common` states that "buf" should be used instead of
"buffer" but there are cases where buffer is mentioned in the lua API.
Solution:
- Rename occurrences of "buffer" to "buf" for consistency with the
documentation.
- Support (but deprecate) "buffer" for backwards compatibility.
Co-authored-by: Justin M. Keyes <justinkz@gmail.com>
Problem:
Since 2f6d1d3c88, `apply_text_edits`
unconditionally sets `buflisted=true`, causing spurious BufDelete events
if plugins restore the original 'buflisted' state on unlisted buffers:
65ef6cec1c/src/nvim/option.c (L2159-L2169)
Solution:
- Don't set 'buflisted' in `apply_text_edits`. Set it more narrowly, in
`apply_workspace_edit` where the semantics requires affected buffers
to be visible to the user.
- Also skip setting 'buflisted' if it would not be changed, to avoid
redundant `OptionSet` events.
