In the <script> element, enable jsdoc embedding in JavaScript.
* lisp/progmodes/php-ts-mode.el
(php-ts-mode--language-source-alist): Add jsdoc to parser recipe.
(php-ts-mode): Add jsdoc parser check and a new rule for
'treesit-range-settings'. Update the commentary section.
(Bug#71771)
'eldoc-echo-area-use-multiline-p' is there for that.
* lisp/progmodes/python.el (python-eldoc-setup-code):
Make the change. (Bug#71720)
* etc/NEWS: Announce the change in behavior.
* lisp/progmodes/xref.el (xref--xref-buffer-mode-map): Cease
binding 'g' to 'xref-revert-buffer'.
(xref--xref-buffer-mode): Set 'revert-buffer-function' to...
(xref--revert-buffer): ...this. New function, renamed from...
(xref-revert-buffer): ...this. Make it an alias of
'revert-buffer'.
* etc/NEWS: Announce it.
* doc/emacs/maintaining.texi (Xref Commands): Update docs.
* lisp/progmodes/js.el (js--treesit-font-lock-settings): Add jsdoc
font-lock settings.
(js--treesit-jsdoc-beginning-regexp): New variable to match the
beginning of a jsdoc block.
(js-ts-mode): Add jsdoc support. Save the 'javascript parser to
`treesit-primary-parser'. Configure `treesit-range-settings' to
use a jsdoc parser within a JavaScript file.
(js-ts-language-at-point): New function to return either 'jsdoc or
'javascript depending on where the point is.
If the value of a property is text representing a CSS color,
it will be fontified such that its background is the color
itself. 'php-ts-mode-css-fontify-colors' can be used to
disable this behaviour.
* lisp/progmodes/php-ts-mode.el (php-ts-mode-css-fontify-colors):
New custom var.
* lisp/progmodes/php-ts-mode.el (php-ts-mode--colorize-css-value):
New function.
* lisp/progmodes/php-ts-mode.el (php-ts-mode): Use the new
function. (Bug#71724)
The new rule works with both the new (>= 0.22.1) and the
old (<= 0.22.0) grammar libraries.
* lisp/progmodes/c-ts-mode.el (c-ts-mode--keywords): Removed
the keyword "virtual".
* lisp/progmodes/c-ts-mode.el (c-ts-mode--font-lock-settings):
New font lock rule. (Bug#71518)
For some control structures, PHP provides an alternative syntax.
A new rule has been added to handle this syntax.
* lisp/progmodes/php-ts-mode.el (php-ts-mode--indent-styles):
New rule for PHP alternative syntax. (Bug#71710)
The old regex also captured functions with two or more
uppercase characters. This new regex fixes that issue.
* lisp/progmodes/php-ts-mode.el (php-ts-mode--font-lock-settings):
New regex that match only user-defined constants. (Bug#71593)
The Compilation mode recognizes Kotlin/Gradle errors but only in
the now legacy format. The current format, which has been in wide
use for about a year, is not supported. This change adds support
for the current format.
* etc/compilation.txt: (symbols): Add examples of non-legacy
Kotlin/Gradle warnings and errors.
* lisp/progmodes/compile.el: (compilation-error-regexp-alist-alist):
Rename 'gradle-kotlin' to 'gradle-kotlin-legacy' and add
'grade-kotlin' that matches the errors and warnings outputted by
the current (non-legacy) Kotlin/Gradle. (Bug#70797)
* test/lisp/progmodes/compile-tests.el
(compile-tests--test-regexps-data): Rename 'gradle-kotlin' to
'gradle-kotlin-legacy' and add two test cases for the newly added,
non-legacy Kotlin/Gradle warnings and errors.
Prevents php-ts-mode--webserver-read-args from throwing an
error if the current buffer is not associated with a file.
* lisp/progmodes/php-ts-mode.el (php-ts-mode--webserver-read-args):
Use 'default-directory if' the current buffer is not associated
with a file. (Bug#71574)
(php-ts-mode-run-php-webserver): Fix doc string.
* lisp/progmodes/python.el (python-shell-prompt-detect): Use
Python's json package if available, and remove escape sequences
in prompts.
* test/lisp/progmodes/python-tests.el
(python-tests-interpreter-2-6-higher-p): New predicate
function.
(python-shell-prompt-detect-7): New test. (Bug#71440)
* lisp/progmodes/cperl-mode.el (cperl--package-for-imenu-rx)
(cperl--class-for-imenu-rx): avoid false positives when collecting
imenu entries
* lisp/progmodes/cperl-mode.el (cperl-init-faces): avoid false
positives when detecting package names
* test/lisp/progmodes/cperl-mode-tests.el
(cperl-test-imenu-index): Add test verifying that $class_false is
not shown as a class _false
* test/lisp/progmodes/cperl-mode-resources/grammar.pl: Add a line
containing a false positive "$class_false"
* lisp/progmodes/php-ts-mode.el (php-ts-mode-flymake-php):
Don't use `any` to mean `not-newline`; it's misleading.
(php-ts-mode--predefined-constant): Remove duplicate "E_USER_NOTICE".
* etc/NEWS: Announce new features of cperl-mode.
* lisp/progmodes/cperl-mode.el (cperl-menu): Add toggle for extra
paired delimiters.
(defconst): new rx expressions `cperl--sub-name-generated-rx' and
`cperl--field-declaration-rx'
(cperl--imenu-entries-rx): Use the new expressions
(cperl--extra-paired-delimiters): New variable holding the paired
delimiters for Perl 5.36 and newer
(cperl-imenu-sub-keywords): Add autogenerated methods to imenu
(cperl-init-faces): Add the __CLASS__ token, builtin constants,
and attributes for field declarations.
(cperl-short-docs): Add __CLASS__ to one-line docs.
(cperl-extra-paired-delimiters-mode): new minor mode to handle
non-ASCII paired delimiters.
* test/lisp/progmodes/cperl-mode-tests.el
(cperl-test-fontify-builtin-constants): new test.
(cperl-test-fontify-class): New test clauses for attributes.
(cperl-test-field-declaration-rx): new test.
(cperl-test-autogenerated-reader-rx): new unit test for the rx
expression.
(cperl-test-extra-delimiters): new test.
(cperl-test-imenu-index): new clauses for imenu capture of
autogenerated methods.
* test/lisp/progmodes/cperl-mode-resources/grammar.pl: Add
examples for Perl 5.40 syntax.
* test/lisp/progmodes/cperl-mode-resources/extra-delimiters.pl:
New resource for non-ASCII paired delimiters.
* lisp/progmodes/python.el (python-hideshow-forward-sexp-function): Move
point to the end of the line.
* test/lisp/progmodes/python-tests.el (python-hideshow-hide-block-2):
New test. (Bug#71170)
* lisp/progmodes/typescript-ts-mode.el
(typescript-ts-mode--indent-rules): Add indentation rules for
lexical and variable declarations with multiple assignments.
* test/lisp/progmodes/typescript-ts-mode-resources/indent.erts:
Add indent test for variable declarations (bug#68054).
* lisp/progmodes/project.el (project-buffers-viewer): New option.
(project-list-buffers): Dispatch using it.
(project-list-buffers-buffer-menu, project-list-buffers-ibuffer):
New functions (bug#71290).
If the project's root is or destroyed , this may happen for a
variety of reasons, including reasons affecting non-reconnection
scenarios, so better to check in eglot--connect.
An Eglot "reconnection" has always created as accurate a copy of
the previous server object as possible, down to the 'eq'ness or
the saved initargs, language ids, etc. So re-querying for
projects is not the thing to do here.
A new different server can be created with `M-x eglot` which will
ask about shutting down the existing one.
* lisp/progmodes/eglot.el (eglot-reconnect): Revert previous change.
(eglot--connect): Do sanity check here
There are a lot of excluding globs, and checking them all is expensive.
The files glob (i.e. the glob for files we actually want) is usually
just one or two entries, so it's quite fast to check.
If find checks the files glob first and then the excluding glob, it has
to do much less checking (since the files glob will substantially narrow
down the set of files on its own), and find performance is much better.
In my benchmarking, this takes (rgrep "foo" "*.el" "~/src/emacs/trunk/")
from ~410ms to ~130ms.
Further optimizations are possible now that the ignores and matched
files are in the same <F> argument which can be rearranged more easily
without compatibility issues; I'll do those optimizations in later
commits.
* lisp/progmodes/grep.el (rgrep-find-ignores-in-<f>): Add.
* lisp/progmodes/grep.el (rgrep-default-command): Check
rgrep-find-ignores-in-<f> and move the excluded files glob to part of
the "files" argument. (Bug#71179)
* lisp/progmodes/c-ts-common.el: (c-ts-common-comment-indent-new-line):
Single line comment and block comment now behave more like the
c-indent-new-comment-line.
