A recent commit wrapped diff.kak into a module. The module includes the
hook that adds diff highlighting to filetype=diff buffers. This means
that the hook is only loaded after opening the first diff buffer in a
Kakoune session, so it only actually fires for the second diff buffer.
Fix this by moving the hook out of the module.
Fixes#4525
The wrapper for "git blame" creates flags for each line of the buffer.
It parses the output from git and would send a flag (or a series of
flags) each time the commit to blame for a line differs from the
previous one. For files that were touched by a large number of commits,
this results in a high number of kakoune processes being launched, and
may take some time. This is visible in the session through the flags for
the different commits appearing on the lines one by one, possibly during
several seconds.
To speed up the process, batch flags before passing them to the kak
session. One solution could be to send all flags at once, but this might
delay the appearance of commit info for too long if "git blame" really
takes a long time. The alternative solution retained for this commit
consists in grouping as many flags as we can during one second
(roughly), to pass them to kakoune, and then to move on to the next
flags. This way, a new batch of commit information flags appears every
second or so in the client, until all information is added. This should
be much faster than lauching a kakoune process for each commit
reported by "git blame": tests have shown that blaming a large file in
the Linux repository goes 4.5 times faster when batching flags.
Co-authored-by: Johannes Altmanninger <aclopte@gmail.com>
This allows to jump from a mail buffer that contains an inline diff
to the source files (most accurate when the patch has been applied
locally).
This makes the diff module a mandatory dependency; we could relax that.
When reading and writing emails that contain patches (possibly
email-quoted), it can be convenient to the jump to the source file.
Allow this by making diff-jump (bound to <ret> in git-diff
buffers) ignore leading email quotes ("> "). A line that starts with
"> " should not occur in a unified diff, so this won't affect other
use cases.
Observe that diff-jump even works around interleaved replies; they
will not affect the computed line numbers because we ignore lines
that don't match ^(> )*[ +-].
git-diff-goto-source is specific to diffs produced by Git. This patch
generalizes the logic and moves it to a new diff-jump in diff.kak.
The main differences are:
- diff-jump handles plain file diffs (i.e. without the -r option). These
have no "diff" line. This means that it needs to parse +++/--- instead.
- diff-jump can go to the old file, not just the new one.
- diff-jump allows to override the base directory and the number of
directory components to strip.
git-diff-goto-source was implemented with several nested try/catch
blocks. Implementing the extra features would have added more
nesting, redundancy or hidden options. To avoid that, I ported the
parsing logic to Perl (which git.kak already depends on). Maybe
it's possible to do the same in awk.
Potential concerns:
- We could move diff-jump to a new rc/tools/diff.kak but then it's not
obvious where the "diff" module belongs to.
- Should diff "diff-jump -1" be spelled "diff-jump -p1"?
In future, the diff parser could be reused to implement a vimdiff-style
feature: given a diff and the "old" line number, we can compute the
corresponding "new" line number. Perhaps diff-jump should get a -client
argument.
We want to move git-diff-goto-source from rc/tools/git.kak
to rc/filetype/diff.kak (or should we could create
rc/tools/diff.kak?). Either way, create the diff module so we can
formalize this dependency.
Currently this module only provides highlighters, so require it
wherever we reference them.
Keep the diff-select-{file,hunk} commands outside the module because
people might already use them in git buffers.
When `set` or `se` is found at the start of the modeline, it should stop parsing options after `:`.
When `modeline-parse` is called in the following file, it should _not_ recognize `tabstop=4` and `invalid_option=3`.
```
# kak: set indentwidth=0 tabstop=16: tabstop=4 invalid_option=3
```
More info: http://vimdoc.sourceforge.net/htmldoc/options.html#modeline
adds the ability to press <ret> within a hunk and navigate to the original
source code. This can be useful because one often needs to go back and forth
between the diff and the full source code.
- You can press <ret> anywhere _within_ a hunk i.e. lines that start with
` `, `+`, `-`. You will be taken to the exact line in the source that corresponds
to where you pressed <ret> in the hunk. It actually does not make sense
to press <ret> on a `-` line because that does not exist anymore but
in that case you are taken to a nearby line in the hope this is still useful.
- You can also press <ret> on a range line (lines that
look like @@ ... @@). If you press <ret> on anywhere on a range line e.g.
```
@@ -120,3 +123,4 @@ fn some_function {
```
The code will try to navigate to the section heading "fn some_function {"
Note that the section heading is _not_ necessarily located at the
range line (in the above example the range line is 123).
- You can press <ret> on a +++ line also and you will be taken the first
line of the file
Caveats:
- Navigation to the original source file will be accurate only if any edits to
the original source file have been saved to disk, because otherwise
they will not be detected by the `:git diff` or `:git show` commands
- This feature should work well for most typical uses e.g. `:git diff`, `:git diff HEAD^`
`:git diff <some-sha1>`. In fact this feature should work in all scenarios when
the *current files* on disk are being compared _with_ some arbitrary git revision/staging.
It will be less useful in other scenarios when two arbitrary revisions are being
compared to each other or when you are trying to compare staging to some revision.
For example when you invoke `:git diff --staged` you are trying to compare staging
with HEAD but are navigating to what is currently on disk (which may be different
from staging).
Co-authored-by: Johannes Altmanninger <aclopte@gmail.com>
C has header and source files and you need to often switch between them.
Similarly OCaml has .ml (implementation) and .mli (interface files) and
one often needs to switch between them.
This commit provides a simple functionality that allows you to accomplish this.
As per man page eval(1p):
> The eval utility shall construct a command by concatenating arguments together,
> separating each with a `<space>` character. The constructed command shall be
> read and executed by the shell.
When not quoting `$kak_opt_makecmd` in the eval, the variable is split by
newlines and spaces and then joined by spaces to form the command. If there
were newlines in `$kak_opt_makecmd`, the command would be malformed.
To reproduce:
```kak
set-option global makecmd "
echo foo
echo bar"
make a b c
```
Expected output in the `*make*` buffer:
```
foo
bar a b c
```
Actual output:
```
foo echo bar a b c
```
This patch fixes this.
OCaml does not have line comments, and as far as I can tell neither
does Coq. Setting it to '' (like markdown and html do) throws an error
that can be handled or displayed instead of inserting the default '#'.
The closing ``` in the following example was not detected because the
indented code block highlighter was higher up in the hierarchy than the
fenced code block highlighter:
```
indented
```
The codeblock highlighter used to be inline so that it has an effect
inside listblocks. This commits adds a listblock/codeblock highlighter
as a replacement.
Fixes#4351
- Also insert "end" after "do", "else" and "elseif"
- Do not insert "end" after strings or comments containing keywords
- Only insert "end" if the block is empty
An example of the last item is if we want to add a new line to the start
of an unclosed block that already contains statements. @ is the cursor.
-- before
if a then@
x = 1
y = 2
-- after
if a then
@
end
x = 1
y = 2
In this case, inserting the "end" before the statements is probably not
what the programmer wants. It might make more sense to insert the "end"
after the statements, but that is potentially confusing due to spooky
action at a distance. I think the least confusing thing to do in this
situation is to not insert the "end".
This commit makes several improvements to the Lua indentation logic.
- Don't indent if the keyword is inside a string or comment
- Indent inside "do end"
- Indent inside "repeat until"
- Indent after a line ending with "{" or "("
- More accurate un-indentation for the "end" keyword
For the last point, previously we tried to match the indentation of the
starting keyword of the block. However, sometimes this guessed wrong
and produced the wrong indentation, as the following example shows. The
new logic is to indent the "end" by one less level than the contents of
the block.
while true do
if false then
end
end -- This was incorrectly matched with the "if"
kak-lsp uses these faces to mark errors inside the buffer, instead of the Error
face which is much more jarring, and which does not have an associated warning
face. Since the :spell command marks errors inside the buffer, it's also updated
to use this new face.
Adding these faces to Kakoune makes it more likely that colorschemes will
automatically do the right thing when used with kak-lsp, and makes it possible
to use a subtle appearance (like curly underlines) for in-buffer errors while
keeping Kakoune errors bold and jarring as they should be.
rockspec files are used by [Luarocks](https://luarocks.org/), the
most prominent package manager for Lua. Despite the different file
extension, these files are actually Lua files and should be syntax
highlighted as such.
For what it is worth, Neovim also does the same thing that I am doing in
this commit. They recognize both ".lua" and ".rockspec" as being Lua
files (and no other extensions, as far as I know).
This fixes a bug in how the Lua scripts handle new comment lines.
Currently if we have a comment that is indented, when we add a new line
it inserts the `--` prefix before the automatic indentation.
```
--ABC
-- XYZ
```
After the fix, it correctly inserts the comment prefix after the
indentation:
```
--ABC
--XYZ
```
The solution I used is inspired by the ruby.kak script.
If the session wasn't valid anymore by the time the linter finishes,
writing to "$dir"/fifo would hang forever leaving temporary files in
/tmp/kak-lint.XXX and the process alive. This commit fixes that by
not writing to the fifo if the session was not valid.
throwawayaccount12345-1 Copyright Waiver
I dedicate any and all copyright interest in this software to the
public domain. I make this dedication for the benefit of the public at
large and to the detriment of my heirs and successors. I intend this
dedication to be an overt act of relinquishment in perpetuity of all
present and future rights to this software under copyright law.
Correctly indentint on "end" keyword seems very hard,
it is simpler to remove it. And we already insert "end"
in ruby-insert-on-new-line hook, so the removal shouldn't
hurt too much.
currently focus client doesn't work if target pane of the client is in different tab or window from current one.
select window, select tab should be triggered in order to select a pane on a currently not focused tab or window.
1. Deleted back-tick strings which are not a valid Dart syntax
2. Added string interpolation for ${}
3. NOT adding string interpolation for $var as this is considered a bad practice
This commit adds highlighting characters other than the first with red color helping to get immediate feedback of the syntax error.
This helps those who come from languagles like JS where single quotes also denote multi-character strings.
Adding highlighting for string interpolation inside double-quoted strings and character literals inside single-quotes.
Avoiding string interpolation in the form of $var as this is considered bad practice.
The off-by-one was introduced by cd9b1e66 which changed `column-1` to `column`.
The refactoring solves some esoteric quoting errors: I think cases like
unbalanced braces in the bufname and client were not supported.
* bring colors more aligned with upstream solarized implementations
(e.g. vim, emacs)
* tweak rust syntax
* add more details, like operators, highlighting certain traits and
types from std, etc
* remove certain highlighters, like user types. This has the effect
of just highlighting almost the whole code base one color, and
wasn't really correct either. CamelCase for types is only convention
Updates to address these comments:
"I still dont quite see why we need to introduce all those options, especially with names that dont say anything about kotlin. I would expect a single kotlin_static_words option to be enough.
Similarly, a single highlighter should be enough here."
This should cover all filetypes that already auto-insert comments,
except for rust.kak, which is left for a follow-up.
Most of these are straightforward, some explanation for special cases:
rc/filetype/zig.kak rc/filetype/cue.kak
These indent hooks used their own logic to indent after "{" only if
no comment was inserted. Replace this logic by checking if a comment
was inserted. This works because these "*-insert" hooks are run
before their respective "*-indent" hooks.
rc/filetype/php.kak
This also has some logic to insert "*" after "/*" lines. Basic
usage seems to work still. In future this should borrow from the
c-family one, which works a bit better.
Function taking a parameter with a struct tag on the last line
before the opening { were wrongly treated as structs. Add some
additional regex logic to try to catch those cases.
Fixes#4136
The invalid regex `)\b` currently matches anything, so this didn't cause
any errors.
It is still invalid though, so I fixed it by moving the `\b` to the end
of the non-raw_attribute language name (like the original regex). The
raw_attribute one shouldn't need this because the `}` marks the end of
the language name anyway.
Fixes#4025
Current solution makes it difficult to use common Lua practices of having one-liner if statements and using anonymous functions.
New solution prevents auto-indentation and end insertion, if the previous line contains an "end" keyword.
It does not attempt to match each structure with corresponding end, since using multiple end keywords in single line is a very rare occurance in Lua.
Because the HTML highlighter was higher up in the hierarchy than the code
highlighter, it took precedence. I fixed it by making it an inline region.
Using my new knowledge of "inline" I was able to remove one line of code.
Fixes#4091
The X11 repl is unique in that it sends the selection (or parameter) with a
new line appended.
This patch removes that new line and thus, brings it into line with the tmux
and kitty versions.
This commit removes the default prompt value from the `spell-replace`
command.
Currently, running the command after selecting a misspelled word
might not allow the editor to propose alternative spellings because
it completes upon whatever is inserted into the prompt. If the words
returned by `aspell` are too different from the currently misspelled
word, no candidates are shown.
For example, selecting “unanymously” and running `:spell-replace`
will not show any candidates under the current implementation (the
‘y’ probably trips the fuzzy-matcher).
The user develops a habit of clearing the prompt every time, because
that's the only way to make sure all suggestions from `aspell` are
visible, so the editor might as well not have any default value for
`:spell-replace`.
If a line contains three slashes directly followed by a new line, the
next line is also erroneously highlighted as a doc comment currently.
Using a lookahead instead fixes this.
tmux-send-text allows sending an argument, when supplied the argument
will be sent to the REPL instead of the current selection.
tmux-send-text also keeps kak focussed, which does not happen in the x11
variant as it uses xdotool to switch window.
this patch allows:
* Passing an argument to x11-send-text, so that value will be sent
instead of the current selection.
* We capture the window id of the current (presumably kak window)
before we use xdotool to switch window. We can therefore switch back to
kak afterwards (which we do)
This highlighter (line 50 of markdown.kak) looks for the filetype
specified by the author at the top of the code fence, e.g.
``` python
print("hello")
```
and highlights the code within using Kakoune's relevant highlighter --
in this case Python.
Some flavours of markdown use curly braces and other characters in the
first line such as the following:
``` {=python}
print("hello")
```
Previously Kakoune recognised `{=python}` but not `{.python}`. The latter
is Pandoc's flavour of markdown. This patch adjusts the regex patterns
to recognise the dot notation as well.
When $1 or ${kak_selection} start with dash, like "-1", the command will fail, because tmux think it's an argument flag.
-- prevent this.
Also the doc (append new line) is no longer valid.
Other scripts uses a dot `.` to separate the seed from the rest of
the template, making that standard across the codebase allows running
cleanup commands like `rm -rf /tmp/kak-*.*`.
This commit is an attempt at mitigating stray processes and temporary
directories, which pile up in the process tree and `$TMPDIR` over time.
To reproduce the issue, run the `lint` command in rapid successions,
or simply run `:lint; lint; lint;` in the prompt (two consecutive
calls are enough to trigger the bug).
The first call creates a `\*lint-ouput*` buffer, bound to a named
pipe that will be populated later on in an asychronous shell
process. It's that same process that runs the linter afterward, and as
soon as it has been spawned, the following call to `:lint` is executed.
Each call to `:lint` overrides the path to the named pipe that was
assigned to `\*lint-output*` by the previous one, resulting in several
asynchronous processes (that write diagnostics to the pipe) hanging
forever — the pipe is never read, and so the process idles.
The command that removes the temporary directory follows the one that
writes to the named pipe, it's never called in the above scenario,
which additionally results in `kak-lint.XXXXXXXX` directories being
left behind in `$TMPDIR`.
(Also) Fixes#3681.
Triple strings are now distinct from docstrings, triple strings
only preceeded by blanks on the line are considered docstrings.
Avoid highlighting of the closing marker using a lookahead, this
is not fully correct as it will break on a double quote triple
docstring containing a single quote triple string but that seems
improbable enough; if we encounter this in the wild we can split
the two docstring formats into separate regions.
This commit prevents the lines following the one that holds the bullet
from being highlighted with the `bullet` face when they're indented:
- The bullet is highlighted properly, so is this sentence
but this line the ones that would follow are not
Fixes#3582
* fix multilines for declarations
* fix names with `-`, which Ninja allows
* fix few cases of `=` operators
* fix reserved keywords in `command = …` right-hand side
Co-authored-by: Frank LENORMAND <1379068+lenormf@users.noreply.github.com>
Now it uses the window id to identify the REPL window. It is stored in
the option x11_repl_id. That way it is possible to have different REPLs
for different buffers or windows.
Fixes issue #2377 by removing the ability to escape backticks in backticks in compliance with asciidoc's behaviour. Adjusted hooks.asciidoc, keys.asciidoc, and command-parsing.asciidoc to match accordingly.
`-atomic` becomes `-method replace` and `-method overwrite` is now
supported explicitely instead of only available through the
writemethod option.
Fixes#3827
This commit allows code blocks to be prefixed with tabulation
characters to be picked up and highlighted by the editor.
Indenting caused by the inclusion of an inline code block into a
list item is also taken into account. However, that might cause false
positives, for example with a hard wrapped list item indented with
an amount of spaces congruent to 4.
Git includes them in commit messages and notes, which is sometimes useful
when adding shell script snippets:
# this is included in the commit message!
Also, in rebase buffers, # only marks a comment if it is the first nonblank
character in a line.
Remove unnecessary single quotes and whitespaces in %file{}
Do not delete error file before sending to debug buffer
Fix gopls definition error handling and empty sting check
Silence shellcheck warnings
Use the custom object match command for copying indentation of blocks,
rather than simply increasing/decreasing indentation when start and end
statements are encountered.
This fixes an issue where a newline added after an already correctly
placed `else` or `fi` would trigger an unnecessary deindent. Tests have
been added to ensure no regression in this behaviour.
Previously, spelling suggestions were presented with the :menu command,
requiring the user to cycle through wild and fanciful alternatives to get to the
one they wanted. Now, we present suggestions with the :prompt command, which
allows the user to type to filter down the list, and also to customise the
replacement after they've chose it (perhaps to fix capitalisation or add
apostrophe-S).
We also use the mispelled word as the initial content of the prompt. That
filters out the wildest alternatives by default, and allows the user to edit the
original word instead of forcing them to choose from among the suggestions. To
get the full list of suggestions, it's easy enough to just backspace until the
word you want appears in the list.
This adds highlighting for
- quoting operators qw, qr, and qx, like `qw< some words >`
- angle brackets after a quoting operator, like `q<string>`
- punctuation as quoting delimiter, like `q|string|`
- POD sections, which start with ^=\w and and with ^=cut
- heredocs; the marker can be a bare word, or a quoted word, like
print <<~ 'EOF'
single quoted heredoc
EOF
Closes#3736
No attempt is made to use different highlighting for interpolated (qq or
"") strings just yet. Recognizing quote boundaries is more important.
This commit makes `:modeline-parse` grab all lines that look like
modelines, and lets the parser deal with invalid formats.
This allows actually printing an error on unsupported modelines
formats, instead of ignoring them upfront.
This commit prevents specially crafted modelines from making the
editor execute arbitrary Kakoune commands.
By using a tabulation character as a separator, commands can be
injected in the value of the options listed in the modeline. For
example:
# kak: tabstop=2;set-option buffer pwned yes
Fixes#3735.
This commit addresses an off-by-one error that selected `modelines`+1
lines when looking for modelines in a buffer.
Note that setting `modelines` to 0 will still select one line.
Fixes#3733.
