Commit Graph

5652 Commits

Author SHA1 Message Date
Chris Webb
491d4d47ae Fix segfault when adding an invalid default-region highlighter
RegionsHighlighter::create_region() validates the highlighter type argument
but RegionsHighlighter::create_default_region() assumes it is correct,
segfaulting from dereferencing a null pointer if the given type isn't in
the highlighter registry HashMap.

@PJungkamp reported this in https://github.com/mawww/kakoune/issues/4959
with a simple recipe to reproduce:

    :add-highlighter shared/test regions
    :add-highlighter shared/test/ default-region invalid highlighter
    :add-highlighter window/test ref test

Validate the type argument in RegionsHighlighter::create_default_region()
in the same way as RegionsHighlighter::create_region().
2023-08-15 16:22:39 +01:00
Maxime Coste
6942a4c0c9 Change + command not to duplicate identical selections more than once
The current exponential behaviour does not seem that useful, it seems
more predictible that pressing `+` twice would end up with 3 copies
of the original selections instead of 4.

Fixes #4533
2023-08-14 22:50:22 +10:00
Maxime Coste
0a06d9acbd Minor formatting tweaks 2023-08-13 03:57:46 +10:00
Maxime Coste
e090131b87 Add a ProfileScope helper class to replace most profiling uses 2023-08-13 03:57:46 +10:00
Maxime Coste
e605ad8582 Kakoune 2023.08.05 2023-08-05 11:02:11 +10:00
Maxime Coste
d7822ea588 Removed unused captures 2023-08-05 10:39:54 +10:00
Maxime Coste
d1e189f1bf Try to fix clang build issues 2023-08-05 10:34:16 +10:00
Maxime Coste
f58d686066 Kakoune 2023.07.29 2023-07-29 15:53:23 +10:00
Johannes Altmanninger
12310418b0 Allow map/unmap during mapping execution
Commits e49c0fb04 (unmap: fail if the mapping is currently executing,
2023-05-14) 42be0057a (map: fail if key is currently executing,
2023-06-24) fixed potential use-after-free issues. By doing so,
it broke configurations that in practice have not triggered any
crashes [1] [2].

For example with,

	set -remove global autocomplete insert
	hook global InsertCompletionShow .* %{
	    map window insert <esc> <c-o>
	}
	hook global InsertCompletionHide .* %{
	    unmap window insert <esc> <c-o>
	}

The execution of the <esc> mapping triggers InsertCompletionHide fails
at unmapping. This seems legit and I don't see an obvious alternative
way to write it (InsertIdle would not be correct though it would work
in practice).

Fix the regression by allowing map and unmap again while keeping the
mappings alive until they have finished executing.

Applying map/unmap immediately seems like the most obvious semantics.
Alternatively, we could apply them in between key presses.

[1]: <https://github.com/kak-lsp/kak-lsp/issues/689>
[2]: <https://github.com/alexherbo2/auto-pairs.kak/issues/60>
2023-07-20 09:18:23 +02:00
Maxime Coste
e3122ab2c1 Refactor prompt history handling
Share incremental regex logic, pass the synthetized nature of keys
through to input handlers.
2023-07-05 22:00:32 +10:00
Maxime Coste
ec79864559 Merge remote-tracking branch 'krobelus/allow-history-in-mappings' 2023-07-05 20:39:39 +10:00
Maxime Coste
96ff68aeae Merge remote-tracking branch 'krobelus/fix-remap-uaf' 2023-07-04 19:33:10 +10:00
Maxime Coste
53fed4b8b9 Only auto-insert completion when at the end of the line
Auto inserting in the middle is annoying more often than not.
2023-07-04 17:15:33 +10:00
Johannes Altmanninger
42be0057a6 map: fail if key is currently executing
If during execution of a mapping, that same mapping is replaced,
there is undefined behavior because we destroy a mapping that we are
still iterating over.

I have been using this mapping inside my kakrc to re-source the kakrc.

	map global user s %{:source "%val{config}/kakrc"<ret>} -docstring 'source "%val{config}/kakrc"'

Now <space>s happens to not trigger undefined behavior because the
mapping stays the same.

However it triggers an assertion added by Commit e49c0fb04 (unmap:
fail if the mapping is currently executing, 2023-05-14), specifically
the destructor of ScopedSetBool that guards mapping execution.

Fix these by banning map of a key that is executing, just like we
did for unmap.

Alternative solution: we could allow mapping (and even unmapping)
keys at any time and keep them alive by moving them into a trash can,
like we do for clients and others.
2023-07-03 19:03:11 +02:00
Maxime Coste
661d1a0905 Merge common docstring in key mapping assistant
Fixes #4942
2023-07-03 20:48:59 +10:00
Maxime Coste
4b605c582c Merge remote-tracking branch 'omasanori/fixup-undo-doc' 2023-07-03 20:14:21 +10:00
Sergey Fedorov
dca5043812 Unbreak build on ppc
Fixes: https://github.com/mawww/kakoune/issues/4937
2023-06-27 12:29:45 +08:00
Maxime Coste
e06e409dc1 Another small structured binding conversion 2023-06-21 17:36:33 +10:00
Masanori Ogino
58058d1213 Fixup documentation on history navigation commands
Co-authored-by: Johannes Altmanninger <aclopte@gmail.com>
2023-06-21 10:25:59 +09:00
Maxime Coste
e365c42b4b Small structured binding conversion 2023-06-20 19:32:06 +10:00
Maxime Coste
d43268fbeb Fix invalid access of display line end
When a line only contains non-range atoms we can end-up accessing
past the end atom.

Add a test that shows the issue when run with valgrind, it is
unfortunately quite hard to trigger a crash because the invalidly
accessed byte usually leads to the correct code path being taken
(when != DisplayAtom::Range) so we have only 1 in 255 chance of
triggerring a crash.

Fixes #4927
2023-06-20 13:09:03 +10:00
Maxime Coste
e58592f00a Fix highlighters being applied to empty display buffers
In some cases such as with folding we can end-up with regions
not having any atoms to highlight which can trigger a crash as
we assume display buffers not to be empty

Fixes #4926
2023-06-19 12:55:55 +10:00
Johannes Altmanninger
163eb6dbc6 Disable history only for prompts that are never shown in the UI
My terminal allows to map <c-[> and <esc> independently.  I like
to use <c-[> as escape key so I have this mapping:

	map global prompt <c-[> <esc>

Unfortunately, this is not equivalent to <esc>.  Since mappings are
run with history disabled, <c-[> will not add the command to the
prompt history.

So disabling command history inside mappings is wrong in case the
command prompt was created before mapping execution. The behavior
should be: "a prompt that is both created and closed inside a
noninteractive context does not add to prompt history", where
"noninteractive" means inside a mapping, hook, command, execute-keys
or evaluate-commands.

Implement this behavior, it should better meet user expectations.
Scripts can always use "set-register" to add to history.

Here are my test cases:

1. Basic regression test (needs above mapping):

	:nop should be added to history<c-[>

---

2. Create the prompt in a noninteractive context:

	:exec %{:}

now we're back in the interactive context, so we can type:

	nop should be added to history<ret>

---

3. To check if it works for nested prompts, first set up this mapping.

	map global prompt <c-j> '<a-semicolon>:nop should NOT be added to history<ret>'
	map global prompt <c-h> '<a-semicolon>:nop should be added to history first'

Then type

	:nop should be added to history second<c-j><c-h><ret><ret>

the inner command run by <c-j> should not be added to history because
it only existed in a noninteractive context.

---

See also the discussion https://github.com/mawww/kakoune/pull/4692

We could automate the tests if we had a test setup that allowed
feeding interactive key input into Kakoune instead of using
"execute-commands". Some projects use tmux, or maybe we can mock
the terminal.
2023-06-17 11:21:41 +02:00
Johannes Altmanninger
6563b82092 Use auto to avoid repeating type of dynamic cast
I think the clang-tidy lint is called modernize-use-auto
2023-06-17 11:21:16 +02:00
Johannes Altmanninger
00490cd084 Rename "disable_history" stack state to "noninteractive"
The commit after next will fix a bug where we wrongly disable prompt
history in some scenarios. The root cause is that life span of
"disable_history" does not model when we actually want to disable
history.

Let's rename the state variable to "noninteractive". It's set whenever
we are executing a hook, mapping or command.

Note that it's also active inside ":prompt"'s callback, which doesn't
play well with the new name :(
2023-06-17 11:21:16 +02:00
Maxime Coste
5901d2e06b Revert "Switch undo storage from a tree to a plain list"
Moving across history moved to <c-j>/<c-k> to keep <a-u>/<a-U>
for selection undo/redo

This reverts commit e0d33f51b3.
2023-06-17 17:31:57 +10:00
Maxime Coste
b2a853cfc2 Add a few missing entries to the changelog 2023-06-17 17:10:05 +10:00
Maxime Coste
5b1ad0bd0c Add a -previous switch to show-matching highlighter
This switch makes show-matching fallback to the character preceeding
the cursor if the character under the cursor is not a matching
character, which should make show-matching more useful in insert mode.
2023-06-14 22:53:39 +10:00
Maxime Coste
19cbb703a7 Merge remote-tracking branch 'arrufat/fix-build-gcc13' 2023-06-12 21:53:39 +10:00
Maxime Coste
467021c3c7 Merge remote-tracking branch 'arrufat/remove-unneeded-this-capture' 2023-06-12 21:51:13 +10:00
Maxime Coste
ca4593e5cd Fix one missing face pre-parsing 2023-06-12 17:12:24 +10:00
Maxime Coste
6984340936 Store region pointers instead of names in the RegionsHighlighter cache
Cache get fully invalidated whenever the regions change, so there
should be no risk of referencing a removed region, and this removes
one hash map lookup for every region in the displayed buffer range.
2023-06-12 16:45:19 +10:00
Maxime Coste
af66a95ef8 Trim display lines before the colorize pass
Colorizing long lines can be costly, remove all the invisible atoms
earlier. Also optimize ForwardHighlighterApplier further by trimming
empty lines.
2023-06-12 16:26:22 +10:00
Maxime Coste
5a867ebdd1 Pre-parse face specs in Highlighters
Re-parsing face specs can be expensive as highlighters can
be called many times during a redraw with nested regions.
2023-06-10 09:46:46 +10:00
Adrià Arrufat
66b22a1902 Remove unneeded this capture in lambda
Clang 15 was complaining about this.
2023-06-09 21:47:49 +09:00
Adrià Arrufat
77ade51a81 Fix build using GCC 13.1
The <functional> header was missing and the "hash.hh" was not used.
2023-06-09 21:41:51 +09:00
Maxime Coste
e0728d3434 Speed up regions highlighting on files with big lines
Range atoms should always appear in order, so we can iterate a single
time through the display lines and display atoms while applying
hightlighters to regions
2023-06-09 16:38:14 +10:00
Maxime Coste
95a4d70379 Kill current shell on <c-g> during shell execution
Closes #4907
2023-05-29 20:25:56 +10:00
Maxime Coste
cf7c638025 Refactor KeymapManager to enfore setting is_executing on key iteration
Add an iterator based remove to HashMap as that was missing. Make
KeymapManager responsible for throwing on unmap an executing mapping.
2023-05-29 20:11:06 +10:00
Johannes Altmanninger
e49c0fb040 unmap: fail if the mapping is currently executing
When unmapping a key sequence that is currently executing, we continue
executing freed memory which can have weird effects.  Let's instead
throw an error if that happens. In future we can support unmap in
this scenario.

Closes #4896
2023-05-25 00:04:23 +02:00
Maxime Coste
cfa658b899 Add <c-g> to cancel current operation
The current implementation only does this during regex operations,
but should be extensible to other operations that might take a long
time by regularly calling EventManager::handle_urgent_events().
2023-05-21 16:20:51 +10:00
Maxime Coste
e140df8f08 Add an idle callback to be called regularly while regex matching
This paves the way towards being able to cancel long regex matching
operations
2023-05-21 16:20:51 +10:00
Maxime Coste
19b4149d47 Fix warnings with gcc-13 2023-05-21 12:40:27 +10:00
Chris Webb
2f1e536ac7 Fix debug keys output for shift/ctrl modified mouse events
Although Kakoune responds to modified mouse events, they show up in the
debug buffer corrupted. to_string() tests for equality on the mouse event
modifiers rather than testing just the relevant bits, so the modified
mouse events incorrectly fall through to the normal key handling.

Fix this and restructure to allow mouse events to be modifier-prefixed.

Signed-off-by: Chris Webb <chris@arachsys.com>
2023-05-11 20:21:31 +01:00
Maxime Coste
3d989af2de Merge remote-tracking branch 'krobelus/fix-crash-connecting-monitor' 2023-05-10 20:03:22 +10:00
Maxime Coste
459007ae21 Merge remote-tracking branch 'sidkshatriya/remove-ref-ref' 2023-05-10 20:02:51 +10:00
Maxime Coste
d39930cd10 Merge remote-tracking branch 'krobelus/fix-_-at_multibyte_chars' 2023-05-10 19:57:46 +10:00
Chris Webb
fc6f32f9ee Stop _ from tearing multibyte UTF-8 sequences
Fixes #4887

[ja: add test]
2023-05-09 23:07:07 +02:00
Johannes Altmanninger
2308a17833 Update changelog for selection undo 2023-05-09 11:32:37 +02:00
Johannes Altmanninger
7d2bae63b9 Map undo selection change to <a-u>/<a-U>
Change the initial <c-h>/<c-k> bindings to the recently freed-up
<a-u></a-U>.

Pros:
- easier to remember
- the redo binding is logical.
- works on legacy terminals, unlike <c-h>

Cons:
- It's less convenient to toggle between selection undo and redo
  keys. I think this is okay since this scenario does not happen that
  often in practice.
2023-05-09 11:32:37 +02:00