aboutsummaryrefslogtreecommitdiff
path: root/test/functional
Commit message (Collapse)AuthorAge
* fix(marks): ineffective conceal_line callback optimization (#32662)luukvbaal2025-02-28
| | | | | | | Problem: _on_conceal_line callbacks are not invoked if callback has not let Nvim know it wants to receive them. But this may change on factors other than what is currently checked (changed buffer). Solution: Forego this optimization, callback is still guarded behind 'conceallevel'.
* vim-patch:8.2.4603: sourcing buffer lines is too complicatedzeertzjq2025-02-28
| | | | | | | | | | | | | | | | | | | | | | | | Problem: Sourcing buffer lines is too complicated. Solution: Simplify the code. Make it possible to source Vim9 script lines. (Yegappan Lakshmanan, closes vim/vim#9974) https://github.com/vim/vim/commit/85b43c6cb7d56919e245622f4e42db6d8bee4194 This commit changes the behavior of sourcing buffer lines to always have a script ID, although sourcing the same buffer always produces the same script ID. vim-patch:9.1.0372: Calling CLEAR_FIELD() on the same struct twice Problem: Calling CLEAR_FIELD() on the same struct twice. Solution: Remove the second CLEAR_FIELD(). Move the assignment of cookie.sourceing_lnum (zeertzjq). closes: vim/vim#14627 https://github.com/vim/vim/commit/f68517c1671dfedcc1555da50bc0b3de6d2842f6 Co-authored-by: Yegappan Lakshmanan <yegappan@yahoo.com>
* vim-patch:9.1.1155: Mode message not cleared after :silent message (#32667)zeertzjq2025-02-27
| | | | | | | | | | | Problem: Mode message not cleared after :silent message (after 9.0.1634). Solution: Don't reset mode_displayed when the message is empty. (zeertzjq) fixes: neovim/neovim#32641 closes: vim/vim#16744 https://github.com/vim/vim/commit/fce1fa5b618458f6f10028faadc9a9ddc227fe76
* fix(lua): wrong script context for option set by func from nvim_exec2 (#32659)zeertzjq2025-02-27
| | | | | | | | | Problem: Wrong script context for option set by function defined by nvim_exec2 in a Lua script. Solution: Call nlua_set_sctx() after adding SOURCING_LNUM and always set sc_lnum for a Lua script. This is a bug discovered when testing #28486. Not sure if this actually happens in practice, but it's easy to fix and required for #28486.
* fix(display): correctly store winline info for concealed lines (#32656)luukvbaal2025-02-27
| | | Off-by-one error in storing last line number for a logical line.
* fix(popup): reuse pum preview float win, set 'winfixbuf' #32636glepnir2025-02-27
| | | | | | | | | | | Problem: popup floating window is closed and recreated for each item selection, this is a bit wasteful. Solution: - Hide the preview win (instead of closing it) when the menu is still displayed: 1. When selected_item is -1. 2. When switching from an item with an "info" field to one without. - When pum is undisplayed it is still closed.
* fix(move): 'scrolloff' cursor correction no longer handles folds properly ↵luukvbaal2025-02-27
| | | | | | | (#32642) Problem: f58e7d5f passed `&botline` to `plines_win_full()`, (probably) assuming it would be set to the first line of the fold. Solution: Reinstate call to `hasFolding()` to do so.
* test: option set by Lua autocommand has correct script context (#32652)zeertzjq2025-02-27
|
* refactor(shada): fix coverity warning about leaking memory (#32650)zeertzjq2025-02-27
|
* test: unreliable swapfile_preserve_recover_specJustin M. Keyes2025-02-26
| | | | | | | | | | | | | | | | | | | | | | | | Problem: Test sometimes fails on bsd (cirrus ci): test/functional/ex_cmds/swapfile_preserve_recover_spec.lua:589: retry() attempts: 464 test/functional/ex_cmds/swapfile_preserve_recover_spec.lua:590: Expected objects to be the same. Passed in: (table: 0x0d5f1aa44070) { [1] = '' *[2] = '' [3] = '[Process exited 1]' } Expected: (table: 0x0d5ea3eb8718) { [1] = '' *[2] = '[Process exited 1]' [3] = '' } stack traceback: test/testutil.lua:104: in function 'retry' test/functional/ex_cmds/swapfile_preserve_recover_spec.lua:589: in function <test/functional/ex_cmds/swapfile_preserve_recover_spec.lua:556> Solution: Filter out empty items to avoid irrelevant redraw-timing assumptions.
* feat(lua): vim.text.indent()Justin M. Keyes2025-02-26
| | | | | | | | | | | | | Problem: Indenting text is a common task in plugins/scripts for presentation/formatting, yet vim has no way of doing it (especially "dedent", and especially non-buffer text). Solution: Introduce `vim.text.indent()`. It sets the *exact* indentation because that's a more difficult (and thus more useful) task than merely "increasing the current indent" (which is somewhat easy with a `gsub()` one-liner).
* fix(shada): ":wshada/:rshada [filename]" with shadafile=NONE #32538IpsumCapra2025-02-26
| | | | | | | | | | | Problem: read/write shada function logic was skipped entirely if it was detected the shadafile option was set to 'NONE'. Solution: The filename is now always resolved. When the shadafile option is set to 'NONE' AND no filename was passed, the filename resolves to an empty string, which causes the read/write functions to return. Regardless of whether the option is set to 'NONE', when a filename is explicitly passed, it gets resolved and the read/write logic is accessed.
* build: move all generator scripts to `src/gen/`Lewis Russell2025-02-26
| | | | | | | | | | | - Move all generator Lua scripts to the `src/gen/` - Add a `.luarc.json` to `src/gen/` - Add a `preload.lua` to `src/gen/` - Add `src` to `package.path` so it aligns with `.luarc.json' - Fix all `require` statements in `src/gen/` so they are consistent: - `require('scripts.foo')` -> `require('gen.foo')` - `require('src.nvim.options')` -> `require('nvim.options')` - `require('api.dispatch_deprecated')` -> `require('nvim.api.dispatch_deprecated')`
* fix(eval): don't shorten $HOME in v:stacktrace (#32634)zeertzjq2025-02-26
|
* fix(lua): don't override script ID from :source (#32626)zeertzjq2025-02-25
| | | | | | | | Problem: When setting an option, mapping etc. from Lua without -V1, the script ID is set to SID_LUA even if there already is a script ID assigned by :source. Solution: Don't set script ID to SID_LUA if it is already a Lua script. Also add _editor.lua to ignorelist to make script context more useful when using vim.cmd().
* fix(move): wrong cursor row on concealed line (#32629)luukvbaal2025-02-25
| | | | Problem: Cursor row calculation does not take into account concealed lines. Solution: Break the loop when the next calculated line is concealed.
* fix(treesitter): nil check query for has_conceal_lineLuuk van Baal2025-02-25
|
* test(ui/cursor_spec): reduce flakiness (#32627)zeertzjq2025-02-25
|
* fix(lsp): resize hover window for concealed linesLuuk van Baal2025-02-25
| | | | | | | | Problem: Height of a (markdown) `vim.lsp.util.open_floating_preview()` window can be reduced to account for concealed lines (after #31324). Solution: Set the window height to the text height of the preview window. Set 'concealcursor' to avoid unconcealing the cursorline when entering the hover window.
* feat(treesitter): vertical conceal support for highlighterLuuk van Baal2025-02-25
| | | | | | | | TSHighlighter now places marks for conceal_lines metadata. A new internal decor provider callback _on_conceal_line was added that instructs the highlighter to place conceal_lines marks whenever the editor needs to know whether a line is concealed. The bundled markdown queries use conceal_lines metadata to conceal code block fence lines.
* feat(marks): add conceal_lines to nvim_buf_set_extmark()Luuk van Baal2025-02-25
| | | | Implement an extmark property that conceals lines vertically.
* test: combined injections (#32611)Artem2025-02-25
| | | | | * refactor: rewrite test without trailing whitespace * test: combined injection tests
* vim-patch:9.1.1145: multi-line completion has wrong indentation for last ↵glepnir2025-02-25
| | | | | | | | | | | | line (#32625) Problem: When expanding omni completion items with newlines (e.g. `then\n\t\nend`), the end statement gets wrong indentation. Solution: Add OPENLINE_FORCE_INDENT flag to make open_line() use second_line_indent directly (glepnir) closes: vim/vim#16614 https://github.com/vim/vim/commit/5090a1fecb86c44be83d55e139ed79b7785fa090
* fix(api): don't override Vimscript SID (#32610)zeertzjq2025-02-25
| | | | | | | | | | Problem: When calling an API from Vimscript to set an option, mapping, etc., :verbose shows that it's set from an API client. Solution: Don't override current_sctx.sc_sid when calling an API from Vimscript. Also fix the inverse case where API channel id is not set when calling an API from RPC. Move channel id into sctx_T to make saving and restoring easier. Related #8329
* fix(lua): SIGSEGV in luv callback with error(nil) #32595phanium2025-02-24
| | | | | | | | | Problem: luv callback `vim.uv.new_timer():start(0, 0, function() error() end)` causes SIGSEGV, since `xstrdup` gets NULL from `lua_tostring`. Similar to: https://github.com/neovim/neovim/commit/a5b1b83a2693ffa7a5a0a22b3693d36ea60051be Solution: Check NULL before `xstrdup`.
* vim-patch:025dc48: runtime(vim): Update base-syntax, match :CompilerSet and ↵zeertzjq2025-02-24
| | | | | | | | | :SynMenu commands (#32605) closes: vim/vim#16713 https://github.com/vim/vim/commit/025dc48e88790133ef0da583b2ce5b9c2232ea9e Co-authored-by: Doug Kearns <dougkearns@gmail.com>
* vim-patch:9.1.1143: illegal memory access when putting a register (#32604)zeertzjq2025-02-24
| | | | | | | | Problem: illegal memory access when putting a register Solution: make sure cursor column doesn't become negative https://github.com/vim/vim/commit/e0029daa3599529d9d438cc51c7ada8580297a39 Co-authored-by: Christian Brabandt <cb@256bit.org>
* feat(complete): CompleteDone reason "cancel", "discard" #32600Evgeni Chasnovski2025-02-23
| | | | | | | Problem: there is no way to distinguish between user's explicit completion stop/cancel and other automated reasons. Solution: update "cancel" reason to be set only on explicit CTRL-e, and set intentionally vague "discard" otherwise.
* fix(float): can set title/footer without setting border #32594Evgeni Chasnovski2025-02-23
| | | | | | | | | Problem: setting title and/or footer without explicitly setting border shows "title/footer/ requires border to be set" error. At the same time, explicitly setting `border = "none"` (which is default) shows expected no-border-no-title-no-footer window without error. Solution: allow setting title/footer without explicitly setting border.
* fix(float): ensure floating window width can fit titleMaria José Solano2025-02-23
|
* fix(lsp): unify get_completion_word for textEdits/insertTextMathias Fussenegger2025-02-22
| | | | | | | | | | | | | | Problem: After https://github.com/neovim/neovim/pull/32377 selecting snippets provided by luals inserted the multi-line text before accepting the candidates. That's inconsistent with servers who provide `textEdit` instead of `insertText` and having lines shift up/down while cycling through the completion candidates is a bit irritating. Solution: Use the logic used for `textEdit` snippets also for `insertText`
* fix(marks): truncate double-width inline virt_text consistently (#32560)zeertzjq2025-02-21
| | | | - Fix wrong cursor position with 'listchars' "precedes". - Always show the '<' truncation character.
* test: adjust multibyte virtual text test (#32557)zeertzjq2025-02-21
| | | | Having more chars after a double-width char makes it easier to spot the bug where truncating it causes the pending chars to be lost.
* fix(treesitter): `TSNode:field()` returns all children with the given fieldRiley Bruins2025-02-21
|
* perf(treesitter): only search for injections within the parse rangeRiley Bruins2025-02-21
| | | | Co-authored-by: Jaehwang Jung <tomtomjhj@gmail.com>
* fix(messages): list_cmd kind for :registers, :au[g] #32531luukvbaal2025-02-20
| | | | | | | | Problem: No kind for `:registers/autocmd/augroup` messages. `:registers` chunks are emitted as separate `msg_show` events. Solution: Add the `list_cmd` kind to the message. Introduce a new `msg_ext_skip_flush` variable to set to true around a group of to be paired message chunks.
* test(completion_spec): make Enter test descriptions more accuratezeertzjq2025-02-21
|
* vim-patch:9.1.1121: Enter does not insert newline with "noselect"glepnir2025-02-21
| | | | | | | | | | | | | | | Problem: Enter does not insert newline with "noselect" when the pum is visible (lifepillar) Solution: When Enter is pressed and no complete-item is selected, ins_compl_prep returns false, and the edit function continues processing Enter to insert a new line. (glepnir) fixes: vim/vim#1653 closes: vim/vim#16653 https://github.com/vim/vim/commit/07f0dbe3aa326fdf4d0f1b1cf7d79df89e91fc6e Co-authored-by: glepnir <glephunter@gmail.com>
* feat(messages): confirm kind for z=, :tselect, inputlist() #32521luukvbaal2025-02-20
| | | | | | | | | | | Problem: Messages preceding a `cmdline_show->prompt` event can not be distinguished as such when receiving the event. (But since `msg_show` handlers should be scheduled, one can already check whether a prompt is active when displaying the message.) Solution: Rather than add a new kind again, use the `confirm` kind. Could be seen as slightly misleading where it is more of a choice rather than a confirmation, but that already applies to `confirm()` as well...
* feat(marks): virtual lines support horizontal scrolling (#32497)zeertzjq2025-02-20
| | | | Add a new field `virt_lines_overflow` that enables horizontal scrolling for virtual lines when set to "scroll".
* fix(keycodes): recognize <Find>, <Select> #28431Mantas Mikulėnas2025-02-20
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | PuTTY sets TERM=xterm, but sends ESC[1~ and ESC[4~ for Home/End keys, which does not match what the 'xterm' terminfo has for khome/kend, so libtermkeys instead reports them as the original DEC VT220 names. The VT220 came with a DEC LK201 keyboard which had the following keys in the area above arrow keys (where PCs now have Ins/Del/Home/End/etc): ┌────────┬────────┬────────┐ │ Find │ Insert │ Re- │ │ │ Here │ move │ ├────────┼────────┼────────┤ │ Select │ Prev │ Next │ │ │ Screen │ Screen │ └────────┴────────┴────────┘ These would send ESC[x~ sequences in the expected order: ┌────────┬────────┬────────┐ │ ESC[1~ │ ESC[2~ │ ESC[3~ │ ├────────┼────────┼────────┤ │ ESC[4~ │ ESC[5~ │ ESC[6~ │ └────────┴────────┴────────┘ Modern terminals continue to use the same sequences for Ins/Del as well as PageUp/PageDn. But the VT220 keyboard apparently had no Home/End, and PuTTY apparently chose to re-purpose the Find/Select key sequences for Home/End (even though it claims to emulate Xterm and this doesn't match what actual Xterm does). So when Home/End are used in Neovim through PuTTY with TERM=xterm (the default setting), libtermkey finds no match for the received sequences in the terminfo database and defaults to reporting them as <Find> and <Select> respectively. PuTTY is not unique here -- tmux *also* sends ESC[1~ and ESC[4~ after its internal translation -- but the difference is that 'tmux' terminfo correctly maps them to Home/End so Neovim recognizes them as such, while PuTTY defaults to using 'xterm' which uses a different mapping. This initial patch only allows Neovim to recognize <Find> and <Select> key codes as themselves, so that the user could manually map them e.g. using ":imap <Find> <Home>". Alternatives: - Using TERM=putty(-256color) would of course be the most correct solution, but in practice it leads to other minor issues, e.g. the need to have different PuTTY config profiles for older or non-Linux systems that lack that terminfo, or tmux's insistence on rendering italics as reverse. - Using Neovim through tmux avoids the problem (as tmux recognizes ESC[1~ on input), but is something that needs to be manually run every time. The keycodes.h constants are slightly misnamed because K_SELECT was already taken for a different purpose.
* fix(treesitter): avoid computing fold levels for empty bufferLuuk van Baal2025-02-19
| | | | | | | | Problem: Computing fold levels for an empty buffer (somehow) breaks the parser state, resulting in a broken highlighter and foldexpr. Cached foldexpr parser is invalid after filetype has changed. Solution: Avoid computing fold levels for empty buffer. Clear cached foldinfos upon `FileType`.
* fix(terminal): avoid more `busy_start` lacking `busy_stop` (#32509)Sean Dewar2025-02-19
| | | | | | | | | | | Problem: after #32458, it may still be possible for `busy_start` UI events to be emitted without matching `busy_stop`s in the terminal. Solution: do `terminal_enter`'s cursor visibility check immediately after setting/restoring State so it occurs before events. This ensures that if pending escape sequences are processed while in `terminal_enter`, the cursor's initial visibility is set before `is_focused` is checked by `term_settermprop`. As a result, we can move the call to `showmode` back to where it was originally.
* test(lua/hl_spec): fix hang on exit with ASAN (#32508)zeertzjq2025-02-18
|
* docs: misc (#32258)dundargoc2025-02-17
| | | | | Co-authored-by: Evgeni Chasnovski <evgeni.chasnovski@gmail.com> Co-authored-by: Julian Visser <12615757+justmejulian@users.noreply.github.com> Co-authored-by: zeertzjq <zeertzjq@outlook.com>
* test(lua/hl_spec): reduce flakiness (#32489)zeertzjq2025-02-17
|
* test: reduce flakiness in highlight tests (#32488)zeertzjq2025-02-17
|
* fix(marks): handle composing in inline virt_text with 'nowrap' (#32477)zeertzjq2025-02-16
|
* fix(marks): handle double-with inline virt_text with 'nowrap' (#32476)zeertzjq2025-02-16
|
* fix(terminal): avoid mismatched `busy_start` without `busy_stop` (#32458)Sean Dewar2025-02-15
| | | | | | | | | | | | | | | | Problem: `showmode` in `terminal_enter` may cause `vpeekc` to process events, which may handle pending escape sequences. If `CSI ? 25 l` is handled to hide the cursor, it may remain hidden even after leaving terminal mode if both `terminal_enter` and (indirectly) `showmode` call `ui_busy_start`, as there is only one matching call to `ui_busy_stop` after leaving terminal mode. Solution: let `terminal_enter` handle setting the initial visibility of the cursor before calling `showmode`. Closes #32456. This simple solution assumes it isn't possible for e.g. `os_breakcheck` to be called indirectly by something else before `terminal_enter` initially handles cursor visibility and after it restores it, which I think is true.
generated by cgit (git 2.34.1) at 2025-08-09 09:18:00 +0000