Problem: Comparing two NULL list fails.
Solution: Change the order of comparing two lists.
7b293c730b
N/A patches for version.c:
vim-patch:8.2.1187: terminal2 test sometimes hangs in the GUI on Travis
Problem: Terminal2 test sometimes hangs in the GUI on Travis.
Solution: Disable Test_zz2_terminal_guioptions_bang() for now.
c85156bb89
vim-patch:8.2.1188: memory leak with invalid json input
Problem: Memory leak with invalid json input.
Solution: Free all keys at the end. (Dominique Pellé, closesvim/vim#6443,
closesvim/vim#6442)
6d3a7213f5
vim-patch:8.2.1196: build failure with normal features
Problem: Build failure with normal features.
Solution: Add #ifdef.
83e7450053
vim-patch:8.2.1198: terminal2 test sometimes hangs in the GUI on Travis
Problem: Terminal2 test sometimes hangs in the GUI on Travis.
Solution: Move test function to terminal3 to see if the problem moves too.
a4b442614c
* LSP: Add support for call hierarchies
* LSP: Add support for call hierarchies
* LSP: Add support for call hierarchies
* LSP: Jump to call location
Jump to the call site instead of jumping to the definition of the
caller/callee.
* LSP: add tests for the call hierarchy callbacks
* Fix linting error
Co-authored-by: Cédric Barreteau <>
* Reuse inccommand preview window
Currently, show_sub (inside ex_substitute) creates a new split on each
run for its existing buffer, and ex_substitute calls close_windows for
it. This functionality seems to relay in delayed operations on window
structures where the close event on the newest window is "cancelled" by
win_grid_alloc. But for multigrid, there is optimization in place in
win_grid_alloc which causes any (unnecessary?) allocations to be
skipped, and thus inccommand preview window is not preserved but closed
immediately. Alternative fix would be to remove said optimization, but
the whole "lets create a new split each time and trash the earlier
window" seems too wasteful.
Fix#11529
* Update failing test
The failing test sets inccommand=split and does `:%s/.`, but isn't
expecting to get any contents for the preview window, other than the
windows status line. Update the test to include the preview window
contents too.
vim-patch:8.2.1054: not so easy to pass a lua function to Vim
vim-patch:8.2.1084: Lua: registering function has useless code
I think I have also opened up the possibility for people to use these
callbacks elsewhere, since I've added a new struct that we should be
able to use.
Also, this should allow us to determine what the state of a list is in
Lua or a dictionary in Lua, since we now can track the luaref as we go.
Autoread now works in TUI too. The checktimestamp test is run at most once every 2 seconds not to poll too much and also because it doesn't make sense on some filesystems. A solution based on filesystem notifications should arrive soon.
* LSP: Add tests & use nvim_buf_get_lines in locations_to_items
This is to add support for cases where the server returns a URI in the
locations that does not have a file scheme but needs to be loaded via a
BufReadCmd event.
* LSP: Don't iterate through all lines in locations_to_items
* fixup! LSP: Don't iterate through all lines in locations_to_items
* fixup! fixup! LSP: Don't iterate through all lines in locations_to_items
* fixup! fixup! fixup! LSP: Don't iterate through all lines in locations_to_items
* eval: fix assertion failure in garbage collection
fixes#12387, #12430
Lists with CopyID+1 linked only from previous_funccal may be removed in
the garbage collection. Therefore, the terms of the assertions are not
correct. This can be confirmed by the following (The l:x with CopyID+1 of
the first function call needs to be removed by garbage collection):
func! s:f()
let l:x = [1]
let g:x = l:
endfunc
for _ in range(2)
call s:f()
endfor
call garbagecollect()
" press any key
* test: add test for #12387, #12430
Problem: 'shortmess' flag "n" not used in two places.
Solution: Make use of the "n" flag consistent. (Nick Jensen, closesvim/vim#6245,
closesvim/vim#6244)
722e505d1a
* lsp: support custom hl groups in show_line_diagnostics
Closes#12472.
* runtime: add docs for the new lsp highlight groups
Co-authored-by: francisco souza <fsouza@users.noreply.github.com>
According to the LSP spec[1], multiple edits can have the same starting
position, and if that is the case, they should be applied in the order
as they come in the array.
The implementation uses a reverse sort to not interfere with non applied
edits, but failed to take into account the spec.
[1] https://microsoft.github.io/language-server-protocol/specifications/specification-3-14/#textedit
* take wrapping into account when computing float height
* factor out size calculation
* add test
* accept and pass through opts.wrap_at in floating_preview
* make padding configurable
* slightly refactor fancy_floating_markdown to make use of make_position
* padding using string.format
* move trim and pad to separate function
* nit
Co-authored-by: Hirokazu Hata <h.hata.ai.t@gmail.com>
* remove mention of backward compat
* make lint happy
Co-authored-by: Hirokazu Hata <h.hata.ai.t@gmail.com>
Old behavior is: foo(${placeholder: bar, ...)
with lots of random garbage you'd never want inserted.
New behavior is: foo(bar, baz)
(which maybe is good, maybe is bad [depends on user], but definitely better than it was).
-----
* Implement rudimentary snippet parsing
Add support for parsing and discarding snippet tokens from the completion items.
Fixes#11982
* Enable snippet support
* Functional tests for snippet parsing
Add simplified real-world snippet text examples to the completion items
test
* Add a test for nested snippet tokens
* Remove TODO comment
* Return the unmodified item if the format is plain text
* Add a plain text completion item
* Change uri_to_fname to not convert non-file URIs
A URI with a scheme other than file doesn't have a local file path.
* fixup! Change uri_to_fname to not convert non-file URIs
* fixup! fixup! Change uri_to_fname to not convert non-file URIs
* fixed hang issue with --headless and -r option specified
Calling the do_more_prompt function in headless mode will freeze neovim because it is eventally in the input-accepting state (the same as waiting for --more--).
* fixed "Press ENTER or type command to continue" to be suppressed
If in headless mode, we need to exit at this point. If we continue, we will enter the normal mode and the message "Press ENTER or type command to continue" will be displayed and we will be in the input waiting state.
* fixed functional ex_cmds tests
* Revert "fixed "Press ENTER or type command to continue" to be suppressed"
This reverts commit a02dc40e3b.
* Revert "fixed functional ex_cmds tests"
This reverts commit 3bdb8da20a.
* fixed conditional again
* added test for fixed hang issue with --headless (#11386)
* add lua function to highlight yanked region
* extract namespace, better naming, default values
* add default for event argument
* free timer
* factor out mark to position calculation
* d'oh
* make sure timer stops before callback (cf. luv example)
* factor out timer, more documentation
* fixup
* validate function argument for schedule
* fix block selection past eol
* correct handling of multibyte characters
* move arguments around, some cleanup
* move utility functions to vim.lua
* use anonymous namespaces, avoid local api
* rename function
* add test for schedule_fn
* fix indent
* turn hl-yank into proper (hightlight) module
* factor out position-to-region function
mark extraction now part of highlight.on_yank
* rename schedule_fn to defer_fn
* add test for vim.region
* todo: handle double-width characters
* remove debug printout
* do not shadow arguments
* defer also callable table
* whitespace change
* move highlight to vim/highlight.lua
* add documentation
* add @return documentation
* test: add check before vim.defer fires
* doc: fixup
If the LSP sends an end line that is larger than what nvim considers to be the
last line, you get an Index out of bounds error when fetching the line from
nvim, a change that was introduced in #12223. This change removes the strict
indexing and checks the return value from nvim_buf_get_lines.
viml/profile: fix missing fixes when merging vim-patch:8.1.0130
Fix#12255
### Steps to reproduce using `nvim -u NORC`
```
nvim -u NORC
:function Test()
:endfunction
:profile start log1
:profile func Test
:call Test()
:profile stop
:profile start log2
:profile func Test
:call Test()
:profile stop
```
### Actual behaviour
#### log1
```
FUNCTION Test()
Called 1 times
...
```
#### log2
```
FUNCTION Test()
Called 2 times
...
```
### Expected behaviour
#### log1
```
FUNCTION Test()
Called 1 times
...
```
#### log2
```
FUNCTION Test()
Called 1 times
...
```
* Make apply_text_edits non-ASCII safe
Use `vim.str_byteindex` to correct starting and ending positions for text edits if the line contains non-ASCII characters.
Fixes#12221
* text_edit may be applied to other buffers
* make sure the buffer is loaded
* add comments
* add test for non-ASCII edits
* lsp: handle kinds not specified in protocol
fix: #12200
If the client set "symbolKind.valueSet",
the client must handle it properly even if it receives a value outside the specification.
* test: add lsp.util.{get_completion_item_kind_name, get_symbol_kind_name} test case
* lsp: make lsp.util.{get_completion_item_kind_name, get_symbol_kind_name} private
Problem: Using closure may consume a lot of memory.
Solution: unreference items that are no longer needed. Add a test. (Ozaki
Kiichi, closesvim/vim#3961)
209b8e3e3b
Besides the special-case in get_scrolloff_value(), it makes sense for
'scrolloff' and 'sidescrolloff' to reflect the correct values (for
plugins, scripts, …).
ref 53d607af9c53accfd634435908fb79061f1212b9
ref #11915
ref #12230
lsp.util.buf_versions must be set in advance.
Use helper.insert to create an anonymous buffer,
so create a named buffer for testing without using insert.
This commit creates 4 new highlight groups:
- LspDiagnosticsErrorSign
- LspDiagnosticsWarningSign
- LspDiagnosticsInformationSign
- LspDiagnosticsHintSign
These highlight groups are linked to their corresponding LspDiagnostics
highlight groups by default.
This lets users choose a different color for their sign columns and
virtualtext diagnostics.
Expose `vim.lsp.buf.diagnostics_by_buf`
This makes it easier to customize the diagnostics behavior. For example
to defer the update they can override the
`textDocument/publishDiagnostics` callback to only call
`buf_diagnostics_save_positions` and then defer the other actions to a
autocmd event.
Offsets of window were not taken into account when sending mouse
coordinates to the terminal. Therefore, when nu or rnu is set, the mouse
coordinates sent to the terminal were not correct. Change it to send the
correct coordinates by subtract window offset from col.
Since "nvim_get_option" is executed on the tui thread as a C function
instead of msgpack-rpc, it accesses global variables that may change on the
main thread.
Problem: A plugin cannot easily expand a command like done internally.
Solution: Add the expandcmd() function. (Yegappan Lakshmanan, closesvim/vim#4514)
80dad48c50
Problem: :echomsg and :echoerr do not handle List and Dict like :echo does.
(Daniel Hahler)
Solution: Be more tolerant about the expression result type.
461a7fcfce
Add lua functional tests for :echo,:echon,:echomsg,:echoerr
because nvim did not port "test_" functions from Vim
that modify internal state.
Testing :echoerr via try/catch is sufficient.
- os.exit(1) is too generic, since code 1 may be caused by Nvim exiting
for some other reason. Change it to os.exit(101).
- style: de-architect json_encode/json_decode calls.
Failure seen in travis macOS job:
https://travis-ci.org/neovim/neovim/jobs/647849133
[ FAILED ] test/functional/plugin/lsp_spec.lua@ 266 SP basic_init test should not send didOpen if the buffer closes before init
test/functional/plugin/lsp_spec.lua:297: exit code
Expected objects to be the same.
Passed in:
(number) 1
Expected:
(number) 0
stack traceback:
test/functional/plugin/lsp_spec.lua:297: in function 'on_exit'
test/functional/plugin/lsp_spec.lua💯 in function 'test_rpc_server'
test/functional/plugin/lsp_spec.lua:272: in function <test/functional/plugin/lsp_spec.lua:266>
Problem: Not restoring Insert mode if leaving a prompt buffer by using a
mouse click.
Solution: Set b_prompt_insert appropriately. Also correct cursor position
when moving cursor to last line.
891e1fd894
Problem: BS in prompt buffer starts new line.
Solution: Do not allows BS over the prompt. Make term_sendkeys() handle
special keys. Add a test.
6b810d92a9
Sloppy code inherited from Vim caused user scripts to be able
to observe the cursor line in an invalid intermediary state,
due to Neovim change callbacks being unbuffered unlike Vim listeners.
Manifested in Vimscript executed from the callback possibly erroring
when `:call`:ing any function,
due to the implicit range `curwin->w_cursor.lnum,curwin->w_cursor.lnum`
failing validation.
Fixed by deferring the call to `changed_lines()` until after
`curwin->w_cursor.lnum` gets its correct value.
Problem: When session-restore creates a terminal buffer with command
like `:edit term://.//16450:/bin/bash`, the buffer gets
a different name (depends on PID). Thus the later call to
`bufexists('term://.//16450:/bin/bash)` will return false.
Solution: Force the buffer name with :file. This as least ensures
the same buffer will show in multiple windows correctly, as
expected when saving the session. But it still has problems:
1. the PID in the buffer name is bogus
2. redundant :terminal buffers still hang around
fix#5250
This makes it possible to restore the working directory of :terminal
buffers when reading those buffers from a session file.
Fixes#11288
Co-authored-by: Justin M. Keyes <justinkz@gmail.com>
It worked for MINGW builds at one point
but it keeps failing now because of perl dependencies
or nvim session issues for tests (named pipes as sockets on Windows?).
Add new "splice" interface for tracking buffer changes at the byte
level. This will later be reused for byte-resolution buffer updates.
(Implementation has been started, but using undocumented "_on_bytes"
option now as interface hasn't been finalized).
Use this interface to improve many edge cases of extmark adjustment.
Changed tests indicate previously incorrect behavior. Adding tests for
more edge cases will be follow-up work (overlaps on_bytes tests)
Don't consider creation/deletion of marks an undoable event by itself.
This behavior was never documented, and imposes complexity for little gain.
Add nvim__buf_add_decoration temporary API for direct access to the new
implementation. This should be refactored into a proper API for
decorations, probably involving a huge dict.
fixes#11598
This is inspired by Atom's "marker index" data structure to efficiently
adjust marks to text insertions deletions, but uses a wide B-tree
(derived from kbtree) to keep the nesting level down.
After PR #8226 an unmapped META key in insert mode behaves like
ESC-<key> (:help i_META).
The behaviour does not fully match, since if <Esc>-<key> is pressed
manually then since it were pressed manually `gotchars` would be called
on the second <key> after insert-mode had already been left.
This would mean that `may_sync_undo` (called from `gotchars`) would
call `u_sync(FALSE)` on the second key (since we would be in normal
mode).
This overall means that <Meta-[something]> behaves differently with
respect to undo than <Esc>[something] when the [something] makes a
change.
As an example, under `nvim -u NONE`:
ihello<M-.>u
leaves the buffer empty, while
ihello<Esc>.u
leaves the buffer with one instance of `hello`.
- Fix by calling u_sync() manually in the new clause under
`normalchar:` in `insert_handle_key`.
- Update test in tui_spec.lua that accidentally relied on the old behaviour.
After cbc8d72fde when editing
the command in the command editing window (q:, q/, q?) it was possible
to switch to the previous tab. Doing so put Nvim in a bad state.
Moreover, switching tabs via the other available mechanisms (gt, gT,
<C-W>gt, <C-W>gT) is not possible when in the command editing window.
Here, the behavior is prevented. It is no longer possible to switch to
the previous tab when editing the command in the command editing window.
The solution is to share code between gt, gT, and g<Tab>. Specifically,
goto_tabpage_lastused now calls through goto_tabpage rather than
directly calling goto_tabpage_tp. Doing so works well because all the
validation enjoyed by gt and gT is present in goto_tabpage.
Problem: When Normal highlight group defines ctermfg/bg, but other
highlight group lacks ctermfg/bg, nvim_get_hl_by_id(hl_id,
v:false) returns -1 for the missing ctermfg/bg instead of just
omitting it.
Solution: checking for -1 in hlattrs2dict()
fix#11680
Problem: The Normal highlight is not defined when compiled with GUI.
Solution: Always define Normal. (Christian Brabandt, closesvim/vim#4072)
f90b6e03a9
Problem: Dict and list could be GC'ed while displaying error in a timer.
(Yasuhiro Matsumoto)
Solution: Block garbage collection when executing a timer. Add
test_garbagecollect_soon(). Add "no_wait_return" to
test_override(). (closesvim/vim#4571)
adc6714aac
In a multi-window scenario, it is possible to return focus to the last
accessed window via n_CTRL-W_p. However, in the case of a multi-tab
scenario, there was previously no way to return focus to the last
accessed *tab*. Here, that ability is added via n_g<tab>.
Additionally, the index of the previous tab is exposed via
tabpagenr('#'), mirroring the existing functionality of winnr('#').
Reduce API surface. We don't need so many variations of functions. Too
many functions means verbose, largely redundant documentation, tests,
and cognitive burden.
Problem: Popupmenu is too far left when completion is long. (Linwei)
Solution: Adjust column computations. (Hirohito Higashi, closesvim/vim#2661)
bb008dd323
Problem: The minimum width of the popup menu is hard coded.
Solution: Add the 'pumwidth' option. (Christian Brabandt, James McCoy,
closesvim/vim#2314)
a8f04aa275
Flaky failure (Travis CI, macOS):
[ RUN ] :terminal (with fake shell) works with gf: 10518.41 ms FAIL
test/functional/terminal/ex_terminal_spec.lua:248: Row 1 did not match.
Expected:
|*^ready $ echo "scripts/shadacat.py" |
|* |
|*[Process exited 0] |
|:terminal echo "scripts/shadacat.py" |
Actual:
|*^ |
|*[Process exited 0] |
|* |
|:terminal echo "scripts/shadacat.py" |
To print the expect() call that would assert the current screen state, use
screen:snapshot_util(). In case of non-deterministic failures, use
screen:redraw_debug() to show all intermediate screen states.
stack traceback:
test/functional/ui/screen.lua:579: in function '_wait'
test/functional/ui/screen.lua:361: in function 'expect'
test/functional/terminal/ex_terminal_spec.lua:248: in function <test/functional/terminal/ex_terminal_spec.lua:245>
Traditionally, when navigating to a specific location from the middle of
the jumplist results in shifting the current location to the bottom of
the list and adding the new location after it. This behavior is not
desireable to all users--see, for example
https://vi.stackexchange.com/questions/18344/how-to-change-jumplist-behavior.
Here, another jumplist behavior is introduced. When jumpoptions (a new
option set added here) includes stack, the jumplist behaves like the
tagstack or like history in a web browser. That is, when navigating to
a location from the middle of the jumplist
2 first
1 second
0 third <-- current location
1 fourth
2 fifth
to a new location the locations after the current location in the jump
list are discarded
2 first
1 second
0 third
<-- current location
The result is that when moving forward from that location, the new
location will be appended to the jumplist:
3 first
2 second
1 third
0 new
If the new location is the same
new == second
as some previous (but not immediately prior) entry in the jumplist,
2 first
1 second
0 third <-- current location
1 fourth
2 fifth
both occurrences preserved
3 first
2 second
1 third
0 second (new)
when moving forward from that location.
It would be desireable to go farther and, when the new location is the
same as the location that is currently next in the jumplist,
new == fourth
make the result of navigating to the new location by jumping (e.g. 50gg)
be the same as moving forward in the jumplist
2 first
1 second
0 third
1 new <-- current location
2 fifth
and simply increment the jumplist index. That change is NOT part of
this patch because it would require passing the new cursor location to
the function (setpcmark) from all of its callees. That in turn would
require those callees to know *before* calling what the new cursor
location is, which do they do not currently.
Having the cursor change column can be surprising.
Force startofline in functional and old tests.
Remove the functional breakindent test, as it's a subset of the oldtest one.
Having the cursor change column can be surprising.
Force startofline in functional and old tests.
Remove the functional breakindent test, as it's a subset of the oldtest one.
- We already find ourselves renaming nvim_execute_lua in tests and
scripts, which suggests "exec" is the verb we actually want.
- Add "exec" verb to `:help dev-api`.
Follow-up to 8969efca8 (Vim patch 8.1.0723)
NOTE: This changes the main entrypoint for running single oldtest files
to not use/require the ".res" extension anymore. But it is handled for
B/C.
Adds a phony rule to run oldtest by filename.
Not going through "$(MAKE)" avoids GNUmakefile being used then (which I
use for WIP things), and it seems like SINGLE_MAKE should be used anyway
probably.
Autocmds may close window while it is being entered, then
win_set_minimal_style(wp) operates on an invalid pointer.
We could silently ignore this instead, but it is unlikely to be
intentional, so it is more useful to show an error.
fix#11383
fixes#11438
Backtrace:
0 schar_from_ascii ( p=0x801cc9e112c3 <error: Cannot access memory at address 0x801cc9e112c3>, c=32 ' ') at ../src/nvim/screen.c:5263
1 0x00007f31460eccc5 in win_line (wp=wp@entry=0x7fffc9df6230, lnum=lnum@entry=11, startrow=startrow@entry=10, endrow=41, nochange=false, number_only=number_only@entry=false) at ../src/nvim/screen.c:4025
2 0x00007f31460eed8e in win_update (wp=wp@entry=0x7fffc9df6230) at ../src/nvim/screen.c:1403
3 0x00007f31460f011f in update_screen (type=<optimized out>) at ../src/nvim/screen.c:502
4 0x00007f3146138ef4 in normal_redraw (s=s@entry=0x7fffd0a5f700) at ../src/nvim/normal.c:1247
5 0x00007f314613b159 in normal_check (state=0x7fffd0a5f700) at ../src/nvim/normal.c:1324
6 0x00007f31460accfe in state_enter (s=0x7fffd0a5f700) at ../src/nvim/state.c:28
7 0x00007f3146143099 in normal_enter (cmdwin=<optimized out>, noexmode=<optimized out>) at ../src/nvim/normal.c:463
8 0x00007f314618b541 in main (argc=<optimized out>, argv=<optimized out>) at ../src/nvim/main.c:580
These options were previously global. A global-local window option
behaves closer to a global option "per default" (i e with :set),
but still supports local behavior via :setl
Also this restores back-compat for nvim_set_option("fcs", ...)
which are currently broken on 0.4.x but worked in earlier versions
- Add vim variable meta accessors: vim.env, vim.{g,v,w,bo,wo}
- Redo gen_char_blob to generate multiple blobs instead of just one
so that multiple Lua modules can be inlined.
- Reorder vim.lua inclusion so that it can use previously defined C
functions and utility functions like vim.shared and vim.inspect things.
- Inline shared.lua into nvim, but also keep it available in runtime.
This covers all "small" inserts and deletes in insert mode, as well
as a few more cases like small normal mode deletes
vim-patch:8.1.0678: text properties as not adjusted for inserted text
NB: this is not the final implementation. Bufhl should be made a
part of the extmark tree, so that "start" adjustment just works
automatically. But "stop" will still need some ad-hoc trickery,
until extended marks natively support ranges (hopefully sooner than
forever).
* Clear 'cc' in nvim_open_win 'minimal' style #11361
Add 'colorcolumn' to the list of options that should be cleared when creating
a 'minimal'-style floating window.
- Use correct implementation of text_edits.
- Send indent options to rangeFormatting and formatting.
- Remove references to vim bindings and filetype from lsp.txt
- Add more examples to docs.
- Add before_init to allow changing initialize_params.
Problem: Using a full path is supported for 'directory' but not for
'backupdir'. (Mikolaj Machowski)
Solution: Support 'backupdir' as well. (Christian Brabandt, closesvim/vim#179)
b782ba475a
Mainly configuration and RPC infrastructure can be considered "done". Specific requests and their callbacks will be improved later (and also served by plugins). There are also some TODO:s for the client itself, like incremental updates.
Co-authored by at-tjdevries and at-h-michael, with many review/suggestion contributions.
We often want to do type checking of public function arguments.
- test: Rename utility_function_spec.lua to vim_spec.lua
- .luacov: Map lua module names
