Problem: Reading files with lines approaching MAXCOL length crashes
with segfault due to colnr_T overflow.
Solution: The split check 'linerest >= MAXCOL' fired too late because
linerest could grow by up to 'size' bytes before the next
check. Change threshold to 'linerest >= MAXCOL - size' to
ensure the line passed to ml_append() stays within colnr_T
range.
Note: supported by AI claude
fixes: vim/vim#17935
closes: vim/vim#18953
closes: vim/vim#193326cc291da06
Co-authored-by: Christian Brabandt <cb@256bit.org>
Problem: ml_append_int() crashes when appending lines near MAXCOL
length due to signed integer overflow in space_needed
calculation.
Solution: Change 'space_needed' from int to long to handle the
'len + INDEX_SIZE' computation without overflow. Update
db_free comparison casts from (int) to (long) to match.
Note: supported by AI claude
related: vim/vim#17935
related: vim/vim#18953
related: vim/vim#193320ece393844
Co-authored-by: Christian Brabandt <cb@256bit.org>
When running out of memory in a libuv callback, try_to_free_memory()
will call mf_release_all(). In this case, mf_sync() cannot call
os_breakcheck() as it'll run the libuv loop recursively, so check
main_loop.recursive to prevent that.
Also fix another possible problem that a terminal buffer may have a
swapfile when encountering an OOM in e.g. terminal_alloc().
Problem: `install_deps.sh` tries to install `ninja` on macOS, but it is
installed on the runners by default, triggering warnings (and wasting
time) on CI.
Solution: Don't `brew install ninja`.
This was added before there was a dedicated language server for
tree-sitter queries. Now that https://github.com/ribru17/ts_query_ls
exists, this is a better option in every regard (and up to now
required manually disabling the builtin linter to avoid duplicate
diagnostics and performance issues).
- Rename ".nvimlog" to "nvim.log"
- doesn't need to be "hidden"/dotfile
- ".log" extension helps with filetype detection
- Also rename "nvim/log" => "nvim/nvim.log"
Define a CMake target for every subdirectory of test/functional that
contains functional tests, and a functionaltest_parallel target that
depends on all those targets, allowing multiple test runners to run in
parallel.
On CI, use at most 2 parallel test runners, as using more may increase
system load and make tests unstable.
Problem:
Some UI clients break if space follows the comma in 'guifont' items.
Also the new 'guifont' default may cause problems if none of the fonts are found.
Solution:
Remove trailing whitespace.
Add a "monospace" fallback which `fontconfig` can resolve on *nix systems.
vim-patch:2e9db8fa9 runtime(doc): add Swedish entries to Make_all.mak, add conversion rule to Makefile
vim-patch:776b07000 installman: Update the sed patterns in installman.sh
vim-patch:9fe5a1d55 runtime(doc): Use ellipses for overlength example comments at :help vim9.txt
vim-patch:75541da4c runtime(doc): Update todo.txt and remove solved issues
vim-patch:9.1.0001: tests: Test_popup_setbuf() fails
vim-patch:4cae20c47 README: Updated version from 91 to 92 in various READMEs
vim-patch:93f05b6d4 Makefile: dosrt target fails because of duplicate lang directories
vim-patch:9.2.0008: MS-Windows: font size calculation may be wrong
vim-patch:9.2.0010: Using Wayland compositor is still slow
vim-patch:9.2.0012: Build fails when wayland is not defined
vim-patch:9.2.0013: parallel make invocation may break Wayland header generation
vim-patch:823f4f12b proto: make proto doesn't handle bool type correctly
Problem: undoing after the prompt is changed breaks it (and causes init_prompt
to abort it and append a new one), as the undo history contains the old prompt.
Solution: like submitting, clear the undo buffer. Don't do it in init_prompt if
the line was empty; that may not result in a new prompt, and causes commands
like "S" to lose the history.
As u_save, etc. wasn't being called by prompt_setprompt, undoing after it fixes
the prompt usually gave undesirable results anyway.
Remove the added undo_spec.lua test, as its approach no longer works as a repro,
and finding a new one seems fiddly.
Problem: prompt_setprompt adjusted cursor col may be negative (<=0 when
1-based), and doesn't check the col of ':
Solution: avoid negative col and adjust correctly if ': col differs from old
prompt's length.
Problem: Implicitly setting message target when 'cmdheight' changes.
Solution: Just use the user configured target. Support "cmd" target
with 'cmdheight' set to 0.
Problem:
When MsgArea highlight is changed, the next message may flash
and disappear because msg_start() renders with stale highlight attributes.
msg_puts_len() uses HL_ATTR(HLF_MSG) to render message text, which
happens before update_screen() calls highlight_changed().
So the message is rendered with outdated attrs.
Solution:
Call highlight_changed() in msg_start().
Problem: Opening a terminal on an empty buffer produces a spurious
buffer update event.
Solution: Don't call deleted_lines_buf() if no lines have been deleted.
Problem: Paging keys being consumed without obvious indicator
in the dialog window can be surprising.
Solution: Display a hint with paging keys in the dialog window title
when paging is active. Recognize <Esc> as mapping to stop
paging.
Problem: hl_blend_attrs() through-blend path only checked HL_UNDERLINE | HL_UNDERCURL
for sp_color blending, missing underdotted, combine/blend attr cache used (a << 16) + b
as map key which could overflow if attr IDs ever exceed 16 bits.
Solution: Use HL_UNDERLINE_MASK in the through-blend path to match all underline
styles. Use uint64_t with (a << 32) | b as map key via HlAttrKey macro so each
attr ID gets its full 32-bit range. Use HLATTRS_INIT in hl_get_underline() to
avoid duplicating default field values.
Problem: prompt_setprompt memory leak/other issues when fixing prompt line for
unloaded buffer, or when ': line number is zero.
Solution: don't fix prompt line for unloaded buffer. Clamp ': lnum above zero.
Problem: vim.on_key() called for each message while cmdline is open.
Cursor is on a seemingly random column when pager is entered.
Entering the pager while the cmdline is expanded can be more
convenient than pressing "g<".
Pager window is unnecessarily clamped to half the shell height.
Setting 'laststatus' while pager is open does not adjust its
dimensions.
Solution: Only call vim.on_key() once when dialog window is opened.
Ensure cursor is at the start of the first message when
entering the pager.
Enter the pager window when "<CR>" is pressed while the
cmdline is expanded.
Don't clamp the pager window height.
Set message windows dimensions when 'laststatus' changes.
Problem: Logic determining messages belonging to the last command to
show with "g<" does not flush pending messages. This can
result in clearing the temporary message history before a
message still belonging to the previous command was emitted.
Solution: Flush pending messages when marking the end of messages
belonging to previous command.
Problem:
stack->offset value is used as an array index.
Solution:
Instead of comparing the stack->offset value to the size in bytes of the array,
compare the actual length of the array.
Problem: A multi-line extmark that ends exactly at a deleted range end
is not invalidated.
Revalidated sign mark is added with incorrect range.
Solution: Remove questionable invalidation range condition which was
originally added to avoid deleting a mark that ends below a
deleted line.
Since splicing for a deleted line, and a replaced range that
explicitly ends at column 0 beyond a deleted line is identical,
we can't try to distinguish these two cases. I.e. :1delete 1
and nvim_buf_set_text(0, 0, 0, 1, 0, {}) yield the same splice
operation.
This means that a multi-line sign_text mark should now span at
least one column beyond its end_row, as seen in the adjusted
test. This is still somewhat unexpected/inconvenient to me
which is what prompted me to try to avoid it with the original
condition.
Add revalidated sign mark back to decor with correct range;
third sign mark added to test exposed a crash.
Problem: A float whose zindex is below the cmdline can be enlarged and
end up behind the cmdline.
Solution: Clamp height of such windows to not extend beyond 'cmdheight'.
Problem: Strange behavior when opening terminal on unloaded buffer.
Solution: For nvim_open_term() ensure the buffer is loaded as it needs
to be read into the terminal. For jobstart() just open the
memfile as the file content isn't needed.
Not going to make nvim_open_term() pass stdin to the terminal when stdin
isn't read into a buffer yet, as other APIs don't read stdin on unloaded
buffer either. There are also other problems with loading buffer before
reading stdin, so it's better to address those in another PR.
