As of recently, shell script candidate completions are computed in
the background, and displayed incrementally.
Sorting completions means we can't show partial results.
We do this for "ctags-search"; but if there is only one ctags file
there is already a sensible order (which is slightly different than
what GNU sort does). So let's preserve the order in that case.
The number of completions is probably too high for an order to be useful.
Similarly, for "man", sorting is probably not very helpful because
there are so many results.
See https://github.com/mawww/kakoune/pull/5035#discussion_r1413015934
The last update is from 2017, it's pretty outdated. Current
support was combined from existing languages such as Ruby and
Zig and tweaked to fit the Pony language.
It turns out diffing was pretty fast, but applying the diff was
sub-optimal as it was constantly inserting/erasing lines which
led to lots of unnecessary shifting. Fix this by manually tracking
a read/write iterator and only shifting when necessary (on keeps,
and inserts).
Today "with-option foo bar command-that-fails" fails with
Error: 1:1: 'evaluate-commands': 1:1: 'with-option': 2:5: 'evaluate-commands': 4:9: 'evaluate-commands': 1:2: 'no-such-command': no such command
but leaks the option value. Fix this by resetting the option and
rethrowing the error. Unfortunately the original stack trace is lost
(questionable behavior inherited from C++?).
The awk-generated highlighters in scheme.kak need '\\.' to obtain '\.'
in the generated kakscript output. Fix the inf/nan rule (which should
generate '(?:inf|nan)\.0') to read '(?:inf|nan)\\.0' in the awk.
Kakoune's balanced strings require that delimiter characters nested inside
them are also paired, so for example in %{ }, each nested { must occur
before a corresponding } to balance it out.
In general this will automatically be the case for code in common scripting
languages, but sometimes regular expressions used for syntax highlighting
do end up containing an unbalanced bracket of one type or another.
This problem is easily solved because there is a free choice of balanced
delimiter characters. However, it can also be worked around by adding
a comment which itself contains an unbalanced delimiter character, to
'balance out' the unpaired one in the regular expression.
These unbalanced comments are not ideal as the semantic role they perform
is easy for a casual reader to overlook. A good example is
catch %{
# indent after lines with an unclosed { or (
try %< execute-keys -draft [c[({],[)}] <ret> <a-k> \A[({][^\n]*\n[^\n]*\n?\z <ret> j<a- gt> >
# indent after a switch's case/default statements
try %[ execute-keys -draft kx <a-k> ^\h*(case|default).*:$ <ret> j<a-gt> ]
# deindent closing brace(s) when after cursor
try %[ execute-keys -draft x <a-k> ^\h*[})] <ret> gh / [})] <ret> m <a-S> 1<a-&> ]
}
in rc/filetype/go/kak. Here, it is not instantly obvious that the comment
containing an unmatched { is required for correctness. If you change the
comment, delete it or rearrange the contents of the catch block, go.kak
will fail to load, and if you cut-and-paste this code as the basis for
a new filetype, it is a loaded gun pointing at your feet.
Luckily, a careful audit of the standard kakoune library turned up only
three such instances, in go.kak, hare.kak and markdown.kak.
The examples in go.kak and hare.kak are easily made robust by replacing
a %{ } with %< > or %[ ] respectively. The example in markdown.kak is
least-intrusively fixed by rewriting the affected regular expression
slightly so it has balanced { and } anyway.
The current handling of preprocessor directives in filetype/c-family.kak
leads to a wall of solid colour for more complicated #if or #define
directives, although #include is already nicely highlighted.
Instead of highlighting an entire directive with the meta face, highlight
just the #define, #if or #elif keyword as meta, treating the rest of the
directive as normal c-family expressions. This significantly improves
the readability of complex macro definitions.
For directives other than #define, #if and #elif, we treat the rest of
the directive as an opaque string in normal face rather than trying to
highlight it, covering cases like #error, #pragma, etc. where the rest
of the line is an error message string or other non-expression content.
This does the right thing for #ifdef and #ifndef too, as we don't highlight
identifiers in c-family text so their arguments should be normal face
anyway.
We already pull in the diff module in mail.kak to enable the nice
:diff-jump behaviour on inline patches. Also enable the shared/diff/
highlighter underneath our shared/mail/ highlighter for inline diffs,
listing it first so mail patterns take precedence over diff patterns.
De-emphasise signatures including the standard '^-- \n' separator in the
same way as quoted text in a reply.
Fixes: https://github.com/mawww/kakoune/issues/4998
In several places, we check for a control character with something like
char c;
[...]
if (c >= 0 and c <= 0x1F)
[...]
When char is signed (e.g. amd64) this is fine, but when char is unsigned
by default (e.g. arm32 and arm64) this generates warnings about the
tautologous check that an unsigned value is non-negative.
Write as
if ((unsigned char) c <= 0x1F)
[...]
which is both correct and not suspicious under both conventions.
In display_buffer.hh, the '->' operator is used on an iterator, but
(surprisingly) this is deprecated from C++20 because of x-value vs
l-value ambiguity. Now clang's -Wdeprecated-declarations warns about it
as we declare -std=c++2a. See
https://www.open-std.org/jtc1/sc22/wg21/docs/papers/2019/p1252r2.pdf
which was adopted for 2019-03.
g++ 13.x is confused by the reinterpret_cast in Kakoune's memory.hh
allocator. Use -Wno-stringop-overflow to silence a large number of
verbose false alarms.
Users who rebind default keys and unmap the originals by binding them
to empty strings with empty docstrings end up with empty lines in the
autoinfo. For example, https://github.com/mawww/kakoune/issues/4918.
Hide completely null bindings which have both an empty mapping and an
empty docstring in the autoinfo, as an easy mechanism for these users to
eliminate the UI noise.
Fixes https://github.com/mawww/kakoune/issues/4918
When using either of
set-option g completers option=my_option
prompt -shell-script-candidates ...
While the search text is empty, the completions will be sorted
alphabetically.
This is bad because it means the most important entries are not listed
first, making them harder to select or even spot.
Let's apply input order before resorting to sorting alphabetically.
In theory there is a more elegant solution: sort candidates (except
if they're user input) before passing them to RankedMatch, and then
always use stable sort. However that doesn't work because we use a
heap which doesn't support stable sort.
Closes#1709, #4813
When doing :write -method replace, make sure we've set the correct mode,
uid and gid on the replacement file before attempting to rename it on
top of the original. This means that the original file is left in place
with correct permissions if anything fails, rather than ending up with
0700 permissions from mkstemp().
When a privileged :write is used with -method replace, it silently resets
the ownership of files to root:root. Restore the original owner and group
in the same way we restore the original permissions. Ownership needs to
be restored before permissions to avoid setuid and setgid bits being set
while the file is still owned by root, and to avoid them being subsequently
lost again on chmod(2).
