Problem: a pattern that involves a backref on a different line does not
match when 'ignorecase' is set (QiWei, after v9.1.0645)
Solution: Use MB_STRNICMP when ignorecase is set, fix tests to close
swapfiles
related: vim/vim#14756fixes: vim/vim#17470closes: vim/vim#18104bf82e58a70
Co-authored-by: Christian Brabandt <cb@256bit.org>
(cherry picked from commit 24020ef2dd)
Problem: potential buffer overrun in bufwrite.c
Solution: Use a temporary variable (John Marriott)
In my Windows 11 Pro 64-bit build MAXPATHL is 1024 and IOSIZE is 1025.
In my Archlinux Linux 64-bit build MAXPATHL is 4096 and IOSIZE is 1025.
In funuction buf_write():
There is a check (line 713) that makes sure the length of fname is less
than MAXPATHL. There is a call to STRCPY() (line 1208) which copies the
string at fname into IObuff (which has size IOSIZE). For Unix builds
fname is set to sfname which may or may not be shorter. However, if
sfname is NULL sfname is set to fname.
Therefore, in builds where MAXPATHL > IOSIZE (eg in my linux build), it
is theoretically possible for the STRCPY() call to exceed the bounds of
IObuff.
This PR addresses this by copying fname into a local variable that has
the same maximum size as fname.
In addition:
Given that the filename is unconditionally overwritten in the for loop,
only copy the directory portion of fname. Move variable i closer to
where it is used.
closes: vim/vim#18095a19b019b87
Co-authored-by: John Marriott <basilisk@internode.on.net>
(cherry picked from commit 4263ec21c2)
Problem: Another outdated comment in eval.c (after 9.1.1665).
Solution: Remove that comment as well. Add a few more tests for mapnew()
that fail without patch 8.2.1672 (zeertzjq).
closes: vim/vim#180896b56711804
(cherry picked from commit 639f9f4cda)
Problem: Outdated comment in eval.c.
Solution: Remove the comment, which is no longer true after 8.2.1672.
Also fix a typo in version9.txt (zeertzjq).
closes: vim/vim#180775d3c39af2a
(cherry picked from commit b9699d5701)
Problem:
with `foldmethod=expr foldexpr=v:lua.vim.treesitter.foldexpr()
foldminlines=0`, deleting lines at the end of the buffer always
reports an invalid top error, because the top value (i.e. the
start line number of the deletion) is always 1 greater than
the total line number of the modified buffer.
Solution:
remove the ml_line_count validation
(cherry picked from commit d73cfefed5)
Problem: v_lock is used when it is not initialized. (Yegappan Lakshmanan)
Solution: Initialize the typval in eval1().
4a091b9978
Co-authored-by: Bram Moolenaar <Bram@vim.org>
(cherry picked from commit 865a28155e)
Problem: on_detach may be called after buf_freeall and other important things,
plus its textlock restrictions are insufficient. This can cause issues such as
leaks, internal errors and crashes.
Solution: disable buffer updates in buf_freeall, before autocommands (like the
order after #35355 and when do_ecmd reloads a buffer). Don't do so in
free_buffer_stuff; it's not safe to run user code there, and buf_freeall already
runs before then; just free them to avoid leaks if buf_freeall autocommands
registered more for some reason.
(cherry picked from commit 2211953266)
Co-authored-by: Sean Dewar <6256228+seandewar@users.noreply.github.com>
Problem:
Buffer-updates on_detach callback is invoked before buf_freeall(), which
deletes autocmds of the buffer (via apply_autocmds(EVENT_BUFWIPEOUT,
...)). Due to this, buffer-local autocmds executed in on_detach (e.g.,
LspDetach) are not actually invoked.
Solution:
Call buf_updates_unload() before buf_freeall().
(cherry picked from commit 285c04e2d0)
Co-authored-by: Jaehwang Jung <tomtomjhj@gmail.com>
Problem: Ctrl-G/Ctrl-T does not ignore the end search delimiter
(irisjae)
Solution: Check if the pattern ends with a search delimiter and ignore
it, unless it is part of the pattern.
fixes: vim/vim#17895closes: vim/vim#17933c03990d30f
Co-authored-by: Christian Brabandt <cb@256bit.org>
(cherry picked from commit 1eca030fb2)
Problem: possible undefined behaviour in mb_decompose(), when using the
same pointer as argument several times
Solution: use separate assignments to avoid reading and writing the same
object at the same time (Áron Hárnási)
closes: vim/vim#17953c43a0614d4
Co-authored-by: Áron Hárnási <aron.harnasi@gmail.com>
(cherry picked from commit 77500c5ad5)
Problem: The :keep{alt,jumps,marks,patterns} commmands are sometimes
misidentified as :k.
Solution: Make sure one_letter_cmd() only returns true for :k and not
other :keep* commands (Doug Kearns).
This currently manifests as missing completion for :keep* commands and
incorrect results from fullcommand().
E.g., fullcommand("keepmarks") returns "k" rather than "keepmarks".
The correct command, however, is executed as command modifiers are
handled specially in do_one_cmd() rather than using find_ex_command().
Fix exists(':k') so that it returns 2 for a full match.
closes: vim/vim#15742ea84202372
Cherry-pick Test_ex_command_completion() from patch 9.1.0624.
Co-authored-by: Doug Kearns <dougkearns@gmail.com>
(cherry picked from commit 70bb7999f7)
When Nvim is started in one terminal emulator,
suspended, and later resumed in a different terminal emulator (as can
happen when using e.g. a multiplexer), the new terminal emulator will
not have all of the same modes enabled that the original terminal
emulator had. This is a problem in particular for in-band resize events
because we were disabling the SIGWINCH signal handler when we determined
that the terminal supported this mode.
However, if the new terminal does not support this mode then the
SIGWINCH handler remains disabled, and Neovim no longer properly
resizes. Instead of disabling the SIGWINCH handler, we track the state
of the resize events mode internally. If the mode is enabled then we
return early from the SIGWINCH handler before performing any ioctls or
other system calls. But if the mode is not enabled we proceed as normal.
(cherry picked from commit b1679f0ab6)
Problem: Patch v8.1.0425 was wrong
Solution: Revert that patch (Hirohito Higashi)
This is because the root cause was fixed in 8.1.0786 and a regression
occurred elsewhere.
related: vim/vim#3455
related: vim/vim#3830fixes: vim/vim#11558closes: vim/vim#178996abe5e4904
Co-authored-by: Hirohito Higashi <h.east.727@gmail.com>
(cherry picked from commit 7230296bdb)
Problem: :bnext doesn't go to unlisted help buffers when cycling
through help buffers (after 9.1.0557).
Solution: Don't check if a help buffer is listed (zeertzjq).
From <https://github.com/vim/vim/issues/4478#issuecomment-498831057>:
> I think we should fix that, since once you get to a non-help buffer
> all unlisted buffers are skipped, thus you won't encounter another
> help buffer.
This implies that cycling through help buffers should work even if help
buffers are unlisted. Otherwise this part of :bnext isn't really useful,
as :h makes help buffers unlisted by default.
related: vim/vim#4478
related: vim/vim#15198closes: vim/vim#179139662f33480
(cherry picked from commit 53ac2ad20a)
Problem:
Cursor is visible in "hidden" floating window.
Solution:
Hide cursor when curwin is a hidden floating window.
Show cursor after returning to a normal (non-hidden) window.
(cherry picked from commit d4f2b9050d)
Co-authored-by: glepnir <glephunter@gmail.com>
Problem: Using `addr` without `range` in nvim_create_user_command gives
"No range allowed" error, inconsistent with `:command -addr` behavior.
Solution: Set EX_RANGE flag when `addr` option is specified to match
`:command` behavior.
(cherry picked from commit 22df649210)
Problem: Since 48e2a736, prompt messages are handled by an actual
active cmdline, resulting in `State` no longer being equal
to `MODE_CONFIRM` which is used in some places. E.g. to
specify the current `mode()` or to re-emit a confirm message.
Solution: Replace `MODE_CONFIRM` with a new `MODE_CMDLINE` sub-mode when
`ccline.one_key/mouse_used` is set. Use it to avoid clearing
mouse_used prompt messages, and to re-emit one_key messages
(when ext_messages is inactive, for which this is unnecessary).
(cherry picked from commit e876a739ee)
Problem: The swapfile attention message is not repeated after clearing
the screen.
After clearing the screen `msg_scrolled` is reset without
clearing other related variables, causing an assert.
Solution: Make the attention message part of the confirm prompt.
Call `msg_reset_scroll()`.
(cherry picked from commit d86d4bacc1)
Problem:
The `nvim_get_keymap()` function is missing the `lhsrawalt` field in its response for mappings with an alternate key representation. This makes its return value inconsistent with its documented `maparg()`-like structure and its formal type definition.
Solution:
Corrects the `keymap_array` function to pass the alternate mapping keys (`current_maphash->m_alt->m_keys`) to `mapblock_fill_dict`. The argument responsible for this was previously hardcoded to `NULL`.
For example, for a mapping of `<C-x>`, the API will now correctly return both `lhsraw` (`<80><fc>^DX`) and `lhsrawalt` (the alternate form, e.g., `^X`).
(cherry picked from commit d523750de0)
Problem: When 'wildmode' is set to include "noselect", the popup menu (pum)
incorrectly retained its scroll position when reopened. This
meant that after scrolling down through the menu with `<C-n>`,
reopening the menu (e.g., by retyping the command and
triggering completion again) would show the menu starting from
the previously scrolled position, rather than from the top.
This could confuse users, as the first visible item would not
be the first actual match in the list.
Solution: Ensure that the popup menu resets its scroll position to the
top when reopened (Girish Palya).
closes: vim/vim#176730cd7f3536b
Co-authored-by: Girish Palya <girishji@gmail.com>
Add a non-static variable for this, otherwise it'll be really hacky.
This avoid having to rewrite many incsearch tests in Lua.
(cherry picked from commit c925e7b8ba)
Problem:
The previous fix in #34314 relies on copying the tree in `tree_root` to
ensure the `TSNode`'s tree cannot be mutated. But that causes the
problem where two calls to `tree_root` return nodes from different
copies of a tree, which do not compare as equal. This has broken at
least one plugin.
Solution:
Make all `TSTree`s on the Lua side always immutable, avoiding the need
to copy the tree in `tree_root`, and make the only mutation point,
`tree_edit`, copy the tree instead.
(cherry picked from commit 168bf0024e)
Problem:
'exrc' files are inherently bound to their location / workspace and
therefore require to "know" their location on the filesystem. However,
currently using `debug.getinfo(1, 'S')` returns `"<nvim>"`.
Solution:
Include the filepath as chunkname in `loadstring()` and `nlua_exec()`.
(cherry picked from commit f7c939fa7a)
Problem: Currently terminal highlight attribute buffers are statically allocated
be the size of `TERM_ATTRS_MAX`. This unique case isn't respected in
some places in the ui_compositor. Due to this, when a terminal window
has lines longer them `TERM_ATTRS_MAX`, the compositor will go past the
end of the buffer causing a crash due to out of bounds access.
Solution: Add check to ensure we don't query terminal highlight attrs
past `TERM_ATTRS_MAX` in `win_line()`.
Fixes#30374
(cherry picked from commit d6d1bfd20d)
Problem: Currently `vim.api.nvim__screenshot()` crashes when called with an
invalid path. This is because we don't check if `fopen()` returns a null
pointer.
Solution: Bail out if `fopen()` returns a null pointer.
Fixes: https://github.com/neovim/neovim/issues/34593
(cherry picked from commit 331de6afa6)
Problem: Topline is preemptively updated by line() in WinResized
autocmd with 'splitkeep' != "cursor".
Solution: Set `skip_update_topline` when 'splitkeep' != "cursor".
(Luuk van Baal)
fe803c8c04
(cherry picked from commit 0b91e9f83b)
Problem: On Windows, the value of `term` is overwritten without freeing
the old allocated value, which may lead to a memory leak.
GCC also gives a "incompatible pointer type" warning about
passing `*term` to os_tty_guess_term().
Solution: Don't override the old allocated value, and copy the guessed
value to `term` if its old value is NULL.
(cherry picked from commit fb5a51e775)
Problem: Currently our CI is failing due to msvc warning 5287. This
warning tells us that we are performing a binary expression with enums
of two different types. In this case however, this is clearly intended
and valid.
Solution: Suppress warning 5287 on the line this occurs.
(cherry picked from commit 0980617c0d)
Problem: When closing a floating window, the next window to be entered
may be unfocusable or hidden.
Solution: Don't enter prevwin when it is unfocusable or hidden. Enter
firstwin instead (like for when prevwin is no longer valid).
(cherry picked from commit 0d658660c2)
This might happen, e.g. if we send a fast RPC reply in a os_breakcheck()
in the middle of redrawing and big ui_ext events are produced.
fixes#31316
(cherry picked from commit b8ee354f12)