Problem: Unable to tell what highlight the prompt part of a
cmdline_show event should have, and whether cmdline_hide was
emitted after aborting.
Solution: Add additional arguments hl_id to cmdline_show, and abort to
cmdline_hide.
It's probably not worth adding the C test files to regular formatting as
they're pretty much never touched, but ensuring the files are formatted
according to our standards and getting rid of warnings is a cheap
one-time fix.
Problem: LSP configs on the runtimepath must have the same name as the
LSP server and must also explicitly set the name in vim.lsp.config. This
is redundant and creates a footgun where a user may accidentally use the
wrong name when assigning to the vim.lsp.config table.
Solution: Return a table from lsp/ runtimepath files instead
CID 516419: Integer handling issues (INTEGER_OVERFLOW)
/src/nvim/message.c: 2242 in msg_puts_display()
2236 }
2237 // Concat pieces with the same highlight
2238 size_t len = maxlen < 0 ? strlen(str) : strnlen(str, (size_t)maxlen);
2239 ga_concat_len(&msg_ext_last_chunk, str, len);
2240 msg_ext_cur_len += len;
2241 // When message ends in newline, reset variables used to format message: msg_advance().
>>> CID 516419: Integer handling issues (INTEGER_OVERFLOW)
>>> Expression "len - 1UL", which is equal to 18446744073709551615, where "len" is known to be equal to 0, underflows the type that receives it, an unsigned integer 64 bits wide.
2242 if (str[len - 1] == '\n') {
2243 msg_ext_cur_len = 0;
2244 msg_col = 0;
2245 }
2246 return;
2247 }
The maximum signed value on 32-bit systems is 2 ^ 31 - 1. When using 2 ^ 31 for
the default timeout, the value would overflow on such systems resulting in
a negative value, which caused a stack trace when calling wait() without
a timeout.
Problem:
`termopen` has long been a superficial wrapper around `jobstart`, and
has no real purpose. Also, `vim.system` and `nvim_open_term` presumably
will replace all features of `jobstart` and `termopen`, so centralizing
the logic will help with that.
Solution:
- Introduce `eval/deprecated.c`, where all deprecated eval funcs will live.
- Introduce "term" flag of `jobstart`.
- Deprecate `termopen`.
Also avoid going down message callstack with empty message, and remove expected grid for some tests where it did not change, and we are just testing for expected messages.
- Replace all uses of vim.regex with simpler Lua patterns.
- Replace all uses of vim.fn.substitute with string.gsub.
- Rework error handling so expected errors are passed back via a return.
- These get routed up an passed to `vim.notify()`
- Any other errors will cause a stack trace.
- Reworked the module initialization of `localfile_arg`
- Updated all type annotations.
- Refactored CLI completion by introduction a parse_cmdline()
function.
- Simplified `show_toc()`
- Refactor highlighting
- Inline some functions
- Fix completion on MacOS 13 and earlier.
- Prefer `manpath -q` over `man -w`
- Make completion more efficient by avoiding vim.fn.sort and vim.fn.uniq
- Reimplement using a single loop
Problem:
Language server version information missing from `:checkhealth vim.lsp`.
Solution:
Store `InitializeResult.serverInfo.version` from the `initialize`
response and display for each client in `:checkhealth vim.lsp`.
Problem: ComplMatchIns doesn't work after multibyte chars
(after v9.1.0936)
Solution: Use (ptr - line) instead of wlv.col (zeertzjq).
closes: vim/vim#16233f4ccada5c3
Problem: cannot highlight completed text
Solution: (optionally) highlight auto-completed text using the
ComplMatchIns highlight group (glepnir)
closes: vim/vim#161736a38aff218
Co-authored-by: glepnir <glephunter@gmail.com>
When a terminal application running inside the terminal emulator sets
the cursor shape or blink status of the cursor, update the cursor in the
parent terminal to match.
This removes the "virtual cursor" that has been in use by the terminal
emulator since the beginning. The original rationale for using the
virtual cursor was to avoid having to support additional UI methods to
change the cursor color for other (non-TUI) UIs, instead relying on the
TermCursor and TermCursorNC highlight groups.
The TermCursor highlight group is now used in the default 'guicursor'
value, which has a new entry for Terminal mode. However, the
TermCursorNC highlight group is no longer supported: since terminal
windows now use the real cursor, when the window is not focused there is
no cursor displayed in the window at all, so there is nothing to
highlight. Users can still use the StatusLineTermNC highlight group to
differentiate non-focused terminal windows.
BREAKING CHANGE: The TermCursorNC highlight group is no longer supported.
Problem: Two separate try/end wrappers, that only marginally differ by
restoring a few variables. Wrappers that don't restore
previous state are dangerous to use in "api-fast" functions.
Solution: Remove wrappers that don't restore the previous state.
Always use TRY_WRAP.
Problem: hard to view an existing buffer in the preview window
Solution: add the :pbuffer command (Yinzuo Jiang)
Similar as `:pedit` and `:buffer` command. `:pbuffer` edits buffer [N]
from the buffer list in the preview window.
`:pbuffer` can also open special buffer, for example terminal buffer.
closes: vim/vim#16222a2a2fe841e
Cherry-pick Test_popup_and_previewwindow_dump() changes from patch
9.0.0625.
Cherry-pick Run_noroom_for_newwindow_test() changes from patches
8.2.0432 and 9.0.0363.
Co-authored-by: Yinzuo Jiang <jiangyinzuo@foxmail.com>
Problem: exclusive selection not respected when re-selecting block mode
(Matt Ellis)
Solution: advance selection by another character when using
selection=exclusive and visual block mode
fixes: vim/vim#16202closes: vim/vim#16219bb95589473
Co-authored-by: Christian Brabandt <cb@256bit.org>
1) move the section at :h inclusive-motion-selection-exclusive a few
lines below, so that it doesn't live in between the 2 exceptions.
2) remove the tag :h :!-range. It's not accurate (because it is actually
a filter) and this command is already described at :h :range!
0a4e57f44a
Co-authored-by: Christian Brabandt <cb@256bit.org>
Problem:
- API functions using `try_start` directly, do not surface the
underlying error message, and instead show generic messages.
- Error-handling code is duplicated in the API impl.
- Failure modes are not tested.
Solution:
- Use `TRY_WRAP`.
- Add tests.
Problem:
Regression from de794f2d24: `vim.diagnostic.setqflist{open=true}` attempts to
open the location list instead of the diagnostics quickfix list if it didn't
exist before. This is because we are using `qf_id` to decide which to open, but
`qf_id=nil` when there is no existing diagnostics quickfix list with a given
title ("Diagnostics" by default).
Solution:
- Revert to using `loclist` to decide which to open.
- Add tests.
Problem:
- API functions using `try_start` directly instead of `TRY_WRAP`, do not
surface the underlying error message, and instead show generic things
like "Failed to set buffer".
- Error handling code is duplicated in the API impl, instead of
delegating to the vim buffer/window handling logic.
Solution:
- Use `TRY_WRAP`.
"wininfo" is going to be my next victim. The main problem with wininfo
is that it is "all or nothing", i e either all state about a buffer in a
window is considered valid or none of it is. This needs to be fixed to
address some long running grievances.
For now this is just a warmup: refactor it from a linked list to a
vector.
## Problem
With 'winfixbuf' enabled, `nvim_win_set_buf` and `nvim_set_current_buf` fail
even if targeting the already-current buffer.
vim.wo.winfixbuf = true
vim.api.nvim_win_set_buf(0, 0)
vim.api.nvim_set_current_buf(0)
Solution:
Check for this condition.
