Problem: One-off error in "count" to make "w_skipcol" zero with
'nosmoothscroll' page scrolling when last virtual line
in a buffer line is exactly the entire window width.
(Hirohito Higashi)
Solution: Properly compute the smallest integer value necessary
to make "w_skipcol" zero (Luuk van Baal)
fixes: #17317closes: #17318
Signed-off-by: Luuk van Baal <luukvbaal@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: buflist_new() leaks ffname and fails to reuse curbuf when
autocommands from buf_freeall change curbuf. Plus, a new
buffer is not allocated in this case, despite what the comment
above claims.
Solution: Remove the condition so ffname is not leaked and so a new
buffer is allocated like before v8.2.4791. It should not be
possible for undo_ftplugin or buf_freeall autocommands to
delete the buffer as they set b_locked, but to stay consistent
with other uses of buf_freeall, guard against that anyway
(Sean Dewar).
Note that buf is set to NULL if it was deleted to guard against the (rare)
possibility of messing up the "buf != curbuf" condition below if a new buffer
happens to be allocated at the same address.
closes: #17319
Signed-off-by: Sean Dewar <6256228+seandewar@users.noreply.github.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: MS-Windows: some minor problems building on AARCH64
Solution: Update Make_cyg_ming with aarch64 specific changes,
document how to build on aarch64 using msys2 packages
closes: #17048
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: Loop that ensures "w_skipcol" is zero with 'nosmoothscroll'
for (half)-page scrolling is inefficient.
Solution: Calculate the required "count" instead of looping until
"w_skipcol" is zero (Luuk van Baal).
fixes: #17301closes: #17306
Signed-off-by: Luuk van Baal <luukvbaal@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: still some problem with the new tutors filetype plugin
Solution: refactor code to enable/disable tutor mode into
tutor#EnableInteractive() function, include a test
(Phạm Bình An)
I find it annoying that Tutor's interactive mode is always on (or debug
mode is off) even when I open a tutor file with :edit command.
I think it makes more sense to make this "interactive mode":
- Always on when it is opened with :Tutor command
- Off otherwise
For more references, see `:help` feature, it is a much better than
:Tutor, since I don't have to run `:let g:help_debug = 1` just to be able
to edit and save a help file
Therefore, I remove `g:tutor_debug`
closes: #17299
Signed-off-by: Phạm Bình An <phambinhanctb2004@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: When a space character is used as a trigger in 'isexpand' option
it doesn't get recognized because skip_to_option_part() skips
spaces after a comma, treating them as option separators
rather than option value (after v9.1.1341)
Solution: manually set the part to a space character (glepnir).
closes: #17305
Signed-off-by: glepnir <glephunter@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: Cannot return to the original text after selecting the next
item when the currently selected item is the last one.
Solution: When continuing to move down past the last item, locate the
original completion at the head/tail nodes of the completed
linked list (glepnir).
closes: #17300
Signed-off-by: glepnir <glephunter@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: When an autocommand executes for a non-current buffer,
'eventignorewin' is only checked from the buffer's last
wininfo (overwrites win_ignore in the loop), not from the
value of 'eventignorewin' in all windows showing the buffer as
described (after v9.1.1084)
Solution: Fix the check and don't use wininfo, as that may only contain
windows that recently showed the buffer. Consider all the
buffer's windows in all tabpages (Sean Dewar).
closes: #17294
Signed-off-by: Sean Dewar <6256228+seandewar@users.noreply.github.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: MS-Windows: error when running evim when space in path of Vim
Solution: properly parse quoted strings (Miguel Barro)
When vim is installed in a path with whitespaces (like
`C:\Program Files (x86)\Vim\vim91\vim.exe`). Launching `evim` or
`vim -d` will try to open the file ` (x86)\Vim\vim91\vim.exe`.
Modern versions of vim simplify shell operation by parsing its own
command line. For example on Linux all vim flavours like `evim`, `rvim`,
etc are symlinks to vim. Then vim uses the `parse_command_name()`
function to identify which version of vim it should launch.
For `evim` if a GUI is available `gvim` is launched for better user
experience. In order to launch `gvim` the original command line is
parsed to be passed *verbatim* to the new instance.
This parsing did not properly handle quoted command lines with
whitespaces such as
```
"C:\Program Files (x86)\Vim\vim91\vim.exe" -y myfile.txt
```
closes: #17295
Signed-off-by: Miguel Barro <miguel.barro@live.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: When 'signcolumn' is set to `number` but a line has a sign
without text, the line number disappears (finite-state-machine)
Solution: Verify that a sign actually contains text before rendering the
line number (glepnir)
fixes: #17169closes: #17282
Signed-off-by: glepnir <glephunter@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: when failing to wipeout a quickfix dummy buffer, it will
remain as a dummy buffer, despite being kept.
Solution: clear its dummy BF_DUMMY flag in this case (Sean Dewar).
closes: #17283
Signed-off-by: Sean Dewar <6256228+seandewar@users.noreply.github.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: heap use-after-free possible when autocommands switch away from the
quickfix dummy buffer, but leave it open in a window.
Solution: close its windows first before attempting the wipe.
(Sean Dewar)
related: #17283
Signed-off-by: Sean Dewar <6256228+seandewar@users.noreply.github.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: Currently, 'smartcase' is respected when completing keywords
using <C-N>, <C-P>, <C-X><C-N>, and <C-X><C-P>. However, when
a user continues typing and the completion menu is filtered
using cached matches, 'smartcase' is not applied. This leads
to poor-quality or irrelevant completion suggestions, as shown
in the example below.
Solution: When filtering cached completion items after typing additional
characters, apply case-sensitive comparison if 'smartcase' is
enabled and the typed pattern includes uppercase characters.
This ensures consistent and expected completion behavior.
(Girish Palya)
closes: #17271
Signed-off-by: Girish Palya <girishji@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: Flag checking logic uses a temporary variable and multiple
bitwise operations in insexpand.c
Solution: Consolidate into a single equality check using bitwise OR and
comparison (glepnir)
closes: #17276
Signed-off-by: glepnir <glephunter@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: CI Tests favor GTK2 over GTK3
Solution: Install GTK3 dependencies and debug packages for CI workflows,
update ASAN suppression list, update required dependency
checks for the tests (Drew Vogel)
closes: #17253
Signed-off-by: Drew Vogel <dvogel@github>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: GTK3 and GTK4 will drop numeric cursor support.
Solution: Adopt GTK3 code and use CSS cursor convention (Drew Vogel).
closes: #14610
Signed-off-by: Drew Vogel <dvogel@github>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: too many strlen() calls in gui.c
Solution: refactor gui.c slightly (John Marriott)
This does the following changes:
- use macro STRCMP() instead of strcmp().
- refactor gui_outstr_nowrap() to remove call to STRLEN().
- refactor get_tabline_label() in attempt to simply it. At the same time
use standard looping construct for iterating over windows in a tab.
Move variables closer to where they are used. Add check that we don't
exceed size of NameBuff.
- small optimisation in get_find_dialog_text() to measure the string
length once.
closes: #17269
Signed-off-by: John Marriott <basilisk@internode.on.net>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: MS-Windows: compile warnings and too many strlen() calls in
os_mswin.c
Solution: refactor os_mswin.c and fix the warning (John Marriott)
This change does the following:
- Fix compile warnings (clang 20.1.3) in `mch_libcall()`:
```
os_mswin.c:902:21: warning: cast from 'MYSTRPROCSTR' (aka 'char *(*)(char *)') to 'MYSTRPROCINT' (aka 'int (*)(char *)') converts to incompatible function type [-Wcast-function-type-mismatch]`
902 | retval_int = ((MYSTRPROCINT)ProcAdd)((LPSTR)argstring);`
| ^~~~~~~~~~~~~~~~~~~~~`
os_mswin.c:914:21: warning: cast from 'MYINTPROCSTR' (aka 'char *(*)(int)') to 'MYINTPROCINT' (aka 'int (*)(int)') converts to incompatible function type [-Wcast-function-type-mismatch]`
914 | retval_int = ((MYINTPROCINT)ProcAddI)(argint);`
| ^~~~~~~~~~~~~~~~~~~~~~`
2 warnings generated.
```
- Refactor `stat_impl()` to remove call to `STRLEN()` (via `STRCAT()`).
- Refactor `Messaging_WndProc()`, `enumWindowsGetServer()` and
`serverSendToVim()` to remove calls to `STRLEN()`.
- Use `string_T` to store field `name` in `struct charset_pair` and
`struct quality_pair`. This means we can dispense with some calls to
`STRLEN()`.
- Use `ARRAY_LENGTH()` macro to determine end of array in
`charset_id2name()` and `quality_id2name()` to be consistent with the
rest of the code base.
closes: #17222
Signed-off-by: John Marriott <basilisk@internode.on.net>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: Vim9: type ignored when adding tuple to instance list var
(Lifepillar)
Solution: When getting the typval of class and object member variables,
set the variable type (Yegappan Lakshmanan)
fixes: #17236closes: #17244
Signed-off-by: Yegappan Lakshmanan <yegappan@yahoo.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: [security]: Possible to open more windows into a closing
buffer without splitting, bypassing existing "b_locked_split"
checks and triggering use-after-free
Solution: Disallow switching to a closing buffer. Editing a closing
buffer (via ":edit", etc.) was fixed in v9.1.0764, but add an
error message and check just "b_locked_split", as "b_locked"
is necessary only when the buffer shouldn't be wiped, and may
be set for buffers that are in-use but not actually closing.
(Sean Dewar)
closes: #17246
Signed-off-by: Sean Dewar <6256228+seandewar@users.noreply.github.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: Vim incorrectly escapes tags containing "[" in a help buffer
Solution: check if the buffer has the "help" filetype set, instead of
already being a help buffer (Phạm Bình An)
fixes: #17224closes: #17232
Co-authored-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: zeertzjq <zeertzjq@outlook.com>
Signed-off-by: Phạm Bình An <phambinhanctb2004@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: The pum_redraw function is too complex and difficult to
maintain with nested loops and mixed responsibilities handling
both RTL and LTR text rendering.
Solution: Extracted core rendering logic into dedicated helper functions
(pum_display_rtl_text, pum_display_ltr_text, pum_draw_scrollbar,
pum_process_item) while preserving the original behavior. This
improves code readability and maintainability (glepnir).
closes: #17204
Signed-off-by: glepnir <glephunter@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: tests: Test_terminalwinscroll_topline() fails on Windows
(after v9.1.1348)
Solution: instead of disabling it in Github Actions runners, disable it
for all Windows runs
related: #17196
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: Return value of getcmdline() inconsistent in CmdlineLeavePre
when leaving cmdline in different ways (after v9.1.1329).
Solution: Trigger CmdlineLeavePre before calling abandon_cmdline() so
that getcmdline() can return the command line (zeertzjq).
closes: #17218
Signed-off-by: zeertzjq <zeertzjq@outlook.com>
Signed-off-by: Girish Palya <girishji@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: tests: typo in Test_CmdlineLeavePre_cabbr()
(after v9.1.1349)
Solution: fix typo, disable failing test on Windows for now
(Girish Palya)
closes: #17217
Signed-off-by: Girish Palya <girishji@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: CmdlineLeavePre may trigger twice
(after v9.1.1329)
Solution: check that the key was typed, trigger it when it wasn't before
(Girish Palya)
There are two problems:
- CmdlineLeavePre may be triggered twice when a cabbr is present.
- CmdlineLeavePre fails to trigger when exiting the command-line via
<Backspace>.
Check if the Carriage Return (Enter) key was actually typed.
Trigger the event when the command-line is exited using Backspace and
other keys.
closes: #17214
Signed-off-by: Girish Palya <girishji@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: small problems with gui_w32.c
Solution: fix compile warnings and refactor code (John Marriott)
Compiler (clang v20.1.3) warnings on `_OnMenuSelect()` and
`_OnGetDpiScaledSize()`:
```
clang -c -I. -Iproto -DWIN32 -DWINVER=0x0601 -D_WIN32_WINNT=0x0601
-DHAVE_PATHDEF -DFEAT_HUGE -DHAVE_STDINT_H -D__USE_MINGW_ANSI_STDIO
-pipe -Wall -Wno-deprecated-declarations -D_REENTRANT -U_FORTIFY_SOURCE
-D_FORTIFY_SOURCE=1 -Wall -Wextra -Wshadow -Wstrict-prototypes
-Wmissing-prototypes -Wno-deprecated-declarations
-Wno-error=missing-field-initializers -Werror=uninitialized
-Wunused-but-set-variable -DEXITFREE -DFEAT_GUI_MSWIN -DFEAT_CLIPBOARD
gui_w32.c -o gobjx86-64/gui_w32.o
gui_w32.c:5038:55: warning: comparison of integers of different signs:
'UINT' (aka 'unsigned int') and 'int' [-Wsign-compare]
5038 | && GetMenuState(s_menuBar, pMenu->id, MF_BYCOMMAND) != -1)
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ^ ~~
gui_w32.c:5054:26: warning: unused parameter 'hwnd' [-Wunused-parameter]
5054 | _OnGetDpiScaledSize(HWND hwnd, UINT dpi, SIZE *size)
| ^
2 warnings generated.
```
This commit contains the following changes:
- Fixes Warning 1:
The prototype of `GetMenuState()` says that it returns a UINT, but
returns -1 on failure. Huh?!?
Also, Microsoft says that this function has been superseded (see
https://learn.microsoft.com/en-us/windows/win32/api/winuser/nf-winuser-getmenustate)
and replaced by `GetMenuItemInfo()`. Both of these functions have a
minimum support of Windows 2000.
Therefore in `_OnMenuSelect()`, replace the call to `GetMenuState()`
with `GetMenuItemInfo()`.
- Fixes Warning 2:
Add `UNUSED` to the definition of `_OnGetDpiScaledSize()`.
- Simplify `logfont2name()`.
- Add small optimisations in `_OnNotify()` and `gui_mch_do_spawn()`.
- Add out-of-memory check in `gui_mch_do_spawn()`.
- Code cosmetics (see definitions of `process_message_usual_key_classic()`
and `process_message()`).
closes: #17208
Signed-off-by: John Marriott <basilisk@internode.on.net>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: missing out-of-memory check in textformat.c
Solution: add out-of-memory check, add small optimizations to
internal_format() and same_leader() (John Marriott)
closes: #17200
Signed-off-by: John Marriott <basilisk@internode.on.net>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: tests: Test_xxd_color2() test failure dump diff is misleading
Solution: Ensure the reference dump and the test dump are both processed the
same (Drew Vogel).
Background:
Commit b6dc76b6fd sought to make the xxd
screendump tests compatible with non-standard `XXD` overrides. It provides a
vim script that matches a very general `xxd` command pattern but then also
removed the matching line from the reference dump. This second step was
unnecessary because `VerifyScreenDump()` runs the associated vim script against
both the reference dump and the test dump.
Problem Details:
As part of some unrelated work, the GUI tests were failing with a window size 1
column too narrow. The screendumps in `Test_xxd_color2` were failing as a
result. When I loaded the diff using `term_dumpdiff()` the associated vim
scripts are not run. As a result, the test dump contained the `xxd` invocation
on line 1 while reference dump did not. This throws the diff off, obscuring the
true issue of the GUI window being too narrow.
Verification:
In addition to the test suite, locally I've tried to recreate the issue being
fixed in b6dc76b6fd. Setting a very non-standard
`XXD` path, this test still passes:
```
XXD=/home/dvogel/opt/vim/bin/xxd TEST_FILTER=Test_xxd_color2 TERM=xterm-color make test_xxd
rm -f test_xxd.res test.log messages starttime
if test -n "${ASAN_OPTIONS}"; then \
XXD=../xxd/xxd; export XXD; ASAN_OPTIONS="${ASAN_OPTIONS}_test_xxd" \
UBSAN_OPTIONS="${UBSAN_OPTIONS}_test_xxd" VIMRUNTIME=../../runtime ../vim -f \
-u unix.vim --gui-dialog-file guidialog -U NONE --noplugin --not-a-term -S \
runtest.vim test_xxd.vim ; \
fi
From test_xxd.vim:
Executed Test_xxd_color2() in 0.066049 seconds
Filtered 17 tests with $TEST_FILTER and $TEST_SKIP_PAT
Executed 1 test in 0.096862 seconds
```
closes: #17202
Signed-off-by: Drew Vogel <dvogel@github>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: double free in f_complete_match() (after v9.1.1341)
Solution: remove additional free of trig pointer, correctly free
regmatch.regprog and before_cursor in the error case
closes: #17203
Signed-off-by: glepnir <glephunter@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: Shebang filetype detection can be improved
Solution: Improve detection logic (Eisuke Kawashima)
Vim does not correctly detect filetype from
- `#!/usr/bin/env --split-string=awk -f`
- `#!/usr/bin/env -S -i awk -f`
- `#!/usr/bin/env -S VAR= awk -f`
So update the current detection logic to detect those cases.
closes: #17199
Signed-off-by: Eisuke Kawashima <e-kwsm@users.noreply.github.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
