PROBLEM: The builtin python3 provider cannot auto-detect python3.12
when g:python3_host_prog is not set. As a result, when python3 on $PATH
is currently python 3.12, neovim will fail to load python3 provider
and result in `has("python3") == 0`, e.g.,
"Failed to load python3 host. You can try to see what happened by ..."
ROOT CAUSE: the `system()` call from `provider#pythonx#DetectByModule`
does not ignore python warnings, and `pkgutil.get_loader` will print
a warning message in the very first line:
```
<string>:1: DeprecationWarning: 'pkgutil.get_loader' is deprecated and
slated for removal in Python 3.14; use importlib.util.find_spec() instead
```
SOLUTION:
- Use `importlib.util.find_spec` instead (python >= 3.4)
- Use `-W ignore` option to prevent any potential warning messages
(cherry picked from commit 3da511f1d0)
Problem: r_CTRL-C works differently in visual mode
Solution: Make r_CTRL-C behave consistent in visual mode
in terminal and Windows GUI
in visual mode, r CTRL-C behaves strange in Unix like environments. It
seems to end visual mode, but still is waiting for few more chars,
however it never seems to replace it by any characters and eventually
just returns back into normal mode.
In contrast in Windows GUI mode, r_CTRL-C replaces in the selected area
all characters by a literal CTRL-C.
Not sure why it behaves like this. It seems in the Windows GUI, got_int
is not set and therefore behaves as if any other normal character has
been pressed.
So remove the special casing of what happens when got_int is set and
make it always behave like in Windows GUI mode. Add a test to verify it
always behaves like replacing in the selected area each selected
character by a literal CTRL-C.
closes: vim/vim#13091closes: vim/vim#13112476733f3d0
Co-authored-by: Christian Brabandt <cb@256bit.org>
(cherry picked from commit d2678c78dc)
`LanguageTree:parse` is recursive, and calls
`LanguageTree:for_each_child`, which is also recursive.
That means that, starting from the third level (child of child of root),
nodes will be parsed twice.
Which then means that if the tree is N layers deep, there will be ~2^N
parses even if the branching factor is 1.
Fixes: #25104
This is a maintenance release, focused on bug fixes and improvements.
However, There are included features related to TUI and :terminal.
BRAM
Nvim is a fork of the Vim editor, created and developed by Bram Moolenaar.
On August 3, 2023, he passed away at the age of 62. If Vim or Nvim have
been of use to you in your life, read `:help Bram` and `:help Uganda`
and consider honoring his memory in a way you see fit.
CHANGES SINCE 0.9.1
BREAKING CHANGES
An adjustment was made to the `grid_line` event as part of the exernal
UI protocol the `cells` array might now end with a `[' ', attr, 0]` item
with a repeat count of zero. This is needed by the TUI to disambiguate
final spaces on a line from just clearing the line, which will make a
difference when copying text using the terminal emulators builtin
primary selection support.
External UI:s can safely ignore such an empty item and most UIs already
handle this fine. But it could break some UI:s which has an assert to
validate the cell count to be bigger than zero, or similar checks.
FEATURES
- tui: Support Super and Meta modifiers
- terminal: forward more special keys and modifier-mouse combinations
BUG FIXES
- lua: Always set arg0 to lua scripts
- api: Redundant error when using `nvim_cmd`
- api, lua: Make blank lines in a message work properly
- column: fix bugs related to signs in 'statuscol'
- completion: Don't add backslash in runtime completion
- diff: Filler lines for hunks bigger than linematch limit
- edit: Fix K_EVENT interfering with 'digraph'
- editorconfig: Better validation and error handling
- events: Don't expand non-file as file name
- events: Trigger VimResume on next UI request
- extmarks: Wrong display when changing text with virt_lines
- folds: Update folds in Insert mode with fdm=indent
- helptags: Make multibyte help tags work properly
- highlight: Make CurSearch work properly with 'winhl'
- inccommand: Fix saving of undo info
- keycodes: Recognize <t_xx> as a key
- lsp: Do not assume client capability exists in watchfiles check (#24558)
- mouse: Handle folded lines with virt_lines attached to line above
- remote: Make --remote-expr print to stdout
- remote: Restore previous --remote-expr output formatting
- spell: Splice extmarks on :spellrepall
- startup: Don't truncate when printing with -l
- startup: Run embedded Nvim with real path
- statusline: Redraw when Visual submode changes
- statusline: Fill for double-width char after moving items
- treesitter: updates to queries and injections
- treesitter: Fix TSNode:tree() double free
- ui: Propagate line wrapping state on grid_line events
- ui: Avoid ambiguity about chunk that clears part of line
PERFORMANCE
- extmarks: Avoid unnecessary marktree traversal with folds
- substitute: Don't reallocate new_start every time
BUILD SYSTEM
- deps: Bump libvterm to 0.3.3
- deps: Bump LuaJIT to HEAD - 03c31124c
- deps: Bump libuv to v1.46.0
- deps: Bump Luv to 1.45.0-0
- deps: Bump tree-sitter-c to v0.20.5
- deps: Bump tree-sitter-lua to v0.0.18
Problem:
'endofline' can be used to detect if a file ends of <EOL>, however
editorconfig can break this.
Solution:
Set 'endofline' during BufWritePre
Fixes: #24869
(cherry picked from commit 84376afc72)
When an injection has not set include children, make sure not to add
the injection if no ranges are determined.
This could happen when there is an injection with a child that has the
same range as itself. e.g. consider this Makefile snippet
```make
foo:
$(VAR)
```
Line 2 has an injection for bash and a make variable reference. If
include-children isn't set (default), then there is no range on line 2
to inject since the variable reference needs to be excluded.
This caused the language tree to return an empty range, which the parser
now interprets to mean the full buffer. This caused makefiles to have
completely broken highlighting.
Backports #24550.
Adjusts test to expect that didChangeWatchedFiles is *not* registered if
client `capabilities` is nil (as it's not enabled by default for v0.9).
fix(ui): delay win_viewport until screen update #24182
Problem:
Sometimes, when nvim sends the `win_viewport` event, for example when scrolling
with visible folds on the screen, it reports the `scroll_delta` value one batch
into "future". So when the client application is trying to show the new viewport
it's not yet updated, resulting in temporary corruption / screen flickering.
For more details see #23609, and starting from [this comment](
https://github.com/neovide/neovide/pull/1790#issuecomment-1518697747) in
https://github.com/neovide/neovide/pull/1790,, where the issue was first
detected. Note that some of the conclusions in those are not fully accurate, but
the general observations are.
Solution:
When there are pending updates to a Window, delay the `win_viewport` UI event
until the updates are sent. This ensures that there's no flush between sending
the viewport and updating of the lines corresponding to the new viewport.
Document the existing viewport behaviour (for cases where there are no
extra flushes), give a hint about how applications can deal with the slightly
surprising behaviour of the viewport event being sent after the updates.
Fixes https://github.com/neovim/neovim/issues/23609
Problem: bundled `ftplugin/foo/*.vim` are sourced before user
`ftplugin/foo.vim`.
Solution: call `runtime!` once on all patterns to be sourced.
Followup to #23801. Fixes#24003.
(cherry picked from commit 502a7a0558)
fix(fs): make `normalize()` work with '/' path
Problem: Current implementation of "remove trailing /" doesn't
account for the case of literal '/' as path.
Solution: Remove trailing / only if it preceded by something else.
(cherry picked from commit 80ff66118a)
Co-authored-by: Evgeni Chasnovski <evgeni.chasnovski@gmail.com>
fix(codelens): add buffer and line checks before displaying codelens
(cherry picked from commit 928dc33053)
Co-authored-by: Rohit Sukumaran <rohit.sukumaran@kredx.com>
This is a maintenance release, consisting of bug fixes.
- checkhealth: Add shims for health functions (#23241)
- Add vim.lsp.buf.formatting_sync() to deprecated.txt
- treesitter: update parsers and queries
- api: Extmark highlight groups not always included in details
- api: Avoid assertion when autocmd group id is 0
- api: Don't change title when setting buffer in a window
- api: Nvim_get_hl should return default flag
- colorscheme: Try .lua files in 'rtp' before .vim files in 'pp'
- statuscolumn: fix issues when used with virt_lines and signs
- drawline: Make cursorlineopt=screenline work with resized grid
- eval: Prevent double-free in garbage collection
- extmarks: Don't show virt lines for end mark
- float: Make bufpos work properly with resized parent grid
- highlight: Add missing g: prefix for colors_name
- highlight: Combine ColorColumn with low-priority CursorLine
- highlight: Apply 'winblend' to NormalNC
- lua: Inspect_pos respect bufnr when get syntax info
- lua: Vim.split may trim inner empty items
- man.lua: Don't continue on command error
- man.lua: Return support of all sections
- mark: Properly init mark views
- messages: Ensure msg_grid is at top at more prompt
- mkspell: Prevent Unicode character overflow
- mouse: Cmdline click registered as statuscolumn
- mouse: Fix popup menu position check with winbar
- normal: Make "g$" work properly with resized grid
- pum: Show right-click menu above cmdline area
- pum: Position properly with multigrid and floats
- pum: Fix issues with with 'rightleft'
- redo: Make redo of Lua mappings in op-pending mode work
- redraw: Multibyte characters are wrapped at the end of a line
- redraw: Overwrite double-width char with virt_text properly
- ruler: Show ruler of curwin with no statusline in cmdline
- spell: Extmark with spell=false should disable spell
- statusline: Also allow right click when 'mousemodel' is "popup*"
- substitute: Properly check if preview is needed
- termdebug: Handle partial lines passed to callback
- tui: Position cursor at bottom-left before stopping
- tui: Redraw on SIGWINCH even if size didn't change
- tui: Grid_clear properly clears the screen
- tui: Fix title restore with title stack or altscreen
- ui: Send title to newly-attached UI
- windows: Set stdout to binary mode for --api-info
- windows: Revert installation context to per-machine
- build: Include all dependency directories when generating headers
fix(ftplugin): source Lua files after Vimscript files per directory
Problem: Lua ftplugins in runtime take precedence over Vimscript
ftplugins in user configs (even in `after/`).
Solution: Source ftplugins separately per directory, first Vimscript
then Lua.
(cherry picked from commit d2e4386991)
Co-authored-by: Christian Clason <c.clason@uni-graz.at>
Problem: Not enough characters accepted for 'spellfile'.
Solution: Add vim_is_fname_char() and use it for 'spellfile'.
bc49c5f48f
Cherry-pick related doc update from Vim runtime.
Co-authored-by: Bram Moolenaar <Bram@vim.org>
(cherry picked from commit 7e70a1e44b)
fix(lsp): don't register didChangeWatchedFiles when capability not set
Some LSP servers (tailwindcss, rome) are known to request registration
for `workspace/didChangeWatchedFiles` even when the corresponding client
capability does not advertise support. This change adds an extra check
in the `client/registerCapability` handler not to start a watch unless
the client capability is set appropriately.
(cherry picked from commit 78510add5b)
Co-authored-by: Jon Huhn <huhnjon@gmail.com>
fix(lsp): fix relative patterns for `workspace/didChangeWatchedFiles`
(cherry picked from commit 10f102a3a3)
Co-authored-by: Jon Huhn <huhnjon@gmail.com>
Current behaviour of `:Man` is to only work with "number" sections.
This is caused by wrong assumptions about man sections naming.
Also, there was similar assumption about length of section dirs
in `paths` variable.
fixes#23485
Signed-off-by: Vadim Misbakh-Soloviov <git@mva.name>
(cherry picked from commit 209ed16f57)
perf(lsp): load buffer contents once when processing semantic token responses
Using _get_line_byte_from_position() for each token's boundaries was a
pretty huge bottleneck, since that function would load individual buffer
lines via nvim_buf_get_lines() (plus a lot of extra overhead). So each
token caused two calls to nvim_buf_get_lines() (once for the start
position, and once for the end position).
For semantic tokens, we only attach to buffers that have already been
loaded, so we can safely just get all the lines for the entire buffer at
once, and lift the rest of the _get_line_byte_from_position()
implementation directly while bypassing the part that loads the buffer
line.
While I was looking at get_lines (used by _get_line_byte_from_position),
I noticed that we were checking for non-file URIs before we even looked
to see if we already had the buffer loaded. Moving the buffer-loaded
check to be the first thing done in get_lines() more than halved the
average time spent transforming the token list into highlight ranges vs
when it was still using _get_line_byte_from_position. I ended up
improving that loop more by not using get_lines, but figured the
performance improvement it provided was worth leaving in.
(cherry picked from commit dc38eafab5)
Co-authored-by: John Drouhard <john@drouhard.dev>
fix(treesitter): redraw added/removed injections properly
When injections are added or removed make sure to:
- invoke 'changedtree' callbacks for when new trees are added.
- invoke 'changedtree' callbacks for when trees are invalidated
- redraw regions when languagetree children are removed
(cherry picked from commit b68157834a)
Co-authored-by: Lewis Russell <lewis6991@gmail.com>
fix(treesitter): do not calc folds on unloaded buffers
Fixes#23423
(cherry picked from commit 2e08228a16)
Co-authored-by: Lewis Russell <lewis6991@gmail.com>
