Yuan Fu [Sun, 25 Dec 2022 06:08:17 +0000 (22:08 -0800)]
Further generalize treesit-defun functions
Two new functions, treesit-beginning/end-of-thing. And
treesit-thing-at-point's signature changes.
* lisp/treesit.el (treesit-block-type-regexp): New variable.
(treesit-beginning-of-thing)
(treesit-end-of-thing): Generalized from
treesit-beginning/end-of-defun.
(treesit-beginning-of-defun)
(treesit-end-of-defun): Use the new functions.
(treesit-thing-at-point): Accept PATTERN rather than REGEXP and PRED.
(treesit-defun-at-point): Adjust for the new signature of
treesit-thing-at-point.
Yuan Fu [Sun, 25 Dec 2022 04:17:08 +0000 (20:17 -0800)]
Generalize treesit-defun functions to "things"
Change the "defun" in some functions (e.g. treesit--defuns-around) to
"thing". Add a function treesit-thing-at-point.
* lisp/treesit.el (treesit--thing-unpack-pattern): New subroutine.
(treesit-beginning-of-defun)
(treesit-end-of-defun): Use new function treesit--navigate-thing.
(treesit--defuns-around): Generalize into treesit--thing-around.
(treesit--top-level-defun): Generalize into treesit--top-level-thing.
(treesit--navigate-defun): Generalize into treesit--navigate-thing.
(treesit-thing-at-point): Generalized from treesit-defun-at-point.
(treesit-defun-at-point): Use treesit-thing-at-point to do tht work.
* lisp/emacs-lisp/package-vc.el (package-vc-install-dependencies): Add
new function.
(package-vc--unpack-1): Call 'package-vc-install-dependencies' instead
of 'package-compute-transaction' and 'package-download-transaction'.
It is unreasonable to abort the installation, since we cannot expect
all dependencies to be available in the regular archives. Instead we
note which packages couldn't be found, and warn the user that these
will be missing.
Yuan Fu [Sun, 25 Dec 2022 02:59:39 +0000 (18:59 -0800)]
Fix c-ts-mode imenu defun name (bug#60296)
Extract out c-ts-mode--declarator-identifier from
c-ts-mode--fontify-declarator.
* lisp/progmodes/c-ts-mode.el (c-ts-mode--declarator-identifier): New
function.
(c-ts-mode--fontify-defun): Extract out.
(c-ts-mode--defun-name): Use the new function.
Yuan Fu [Sun, 25 Dec 2022 02:24:01 +0000 (18:24 -0800)]
Support treesit-defun-name in tree-sitter major modes
* lisp/progmodes/csharp-mode.el (csharp-ts-mode--defun-name): New
function.
(csharp-ts-mode--imenu-1): Extract into new function.
(csharp-ts-mode): Setup treesit-defun-name-function.
* lisp/progmodes/java-ts-mode.el (java-ts-mode--defun-name): New
function.
(java-ts-mode--imenu-1): Extract into new function.
(java-ts-mode): Setup treesit-defun-name-function.
* lisp/progmodes/js.el (js-treesit-current-defun): Remove function.
This function is not used (for a while already).
(js--treesit-defun-name): New function.
(js--treesit-imenu-1): Extract into new function.
(js-ts-mode): Setup treesit-defun-name-function.
* lisp/progmodes/json-ts-mode.el (json-ts-mode--defun-name): New
function.
(json-ts-mode--imenu-1): Extract into new function.
(json-ts-mode): Setup treesit-defun-name-function.
* lisp/progmodes/python.el (python--treesit-defun-name): New function.
(python--imenu-treesit-create-index-1): Extract into new function.
(python-ts-mode): Setup treesit-defun-name-function.
* lisp/progmodes/rust-ts-mode.el (rust-ts-mode--defun-name): New
function.
(rust-ts-mode--imenu-1): Extract into new function.
(rust-ts-mode): Setup treesit-defun-name-function.
* lisp/textmodes/css-mode.el (css--treesit-defun-name): New function.
(css--treesit-imenu-1): Extract into new function.
(css-ts-mode): Setup treesit-defun-name-function.
* lisp/textmodes/toml-ts-mode.el (toml-ts-mode--get-table-name):
Remove function.
(toml-ts-mode--defun-name): New function.
(toml-ts-mode--imenu-1): Extract into new function.
(toml-ts-mode): Setup treesit-defun-name-function.
Yuan Fu [Sun, 25 Dec 2022 00:33:35 +0000 (16:33 -0800)]
Add treesit-defun-name and friends
1. We now have treesit-defun-name, powered by
treesit-defun-name-function.
2. We now have treesit-add-log-current-defun, which powers
add-log-current-defun.
3. c-ts-mode updates its code to take advantage of these new features.
4. Manual updates.
* doc/lispref/parsing.texi (Tree-sitter major modes): Add manual for
new functions.
* lisp/progmodes/c-ts-mode.el (c-ts-mode--defun-name): New function.
(c-ts-mode--imenu-1): Extract out into c-ts-mode--defun-name.
(c-ts-base-mode): Setup treesit-defun-name-function.
* lisp/treesit.el (treesit-defun-name-function)
(treesit-add-log-defun-delimiter): New variables.
(treesit-defun-at-point)
(treesit-defun-name): New functions.
(treesit-major-mode-setup): Setup add-log-current-defun-function.
Yuan Fu [Sat, 24 Dec 2022 23:31:03 +0000 (15:31 -0800)]
Make treesit-node-at/on guess language at point
If PARSER-OR-LANG is nil, it makes more sense to guess the language at
point by treesit-language-at than to simply use the first parser in
the parser list.
* doc/lispref/parsing.texi (Retrieving Nodes): Update manual.
* lisp/treesit.el (treesit-node-at)
(treesit-node-on): Guess language at point. Update docstring.
(treesit-buffer-root-node): Update docstring.
Eli Zaretskii [Sat, 24 Dec 2022 19:39:26 +0000 (21:39 +0200)]
Revert "; Bump minimum supported Windows version for MinGW64 to Windows 10."
This reverts commit 75155e458601a3597d382660d0be863ab4d512c0.
Evidently, some MSYS2/MinGW64 folks still think Windows 8.1
is the minimum supported version, even though MinGW-w64 switched
to Windows 10 as the default target in January 2022.
Eli Zaretskii [Sat, 24 Dec 2022 10:08:43 +0000 (12:08 +0200)]
Fix definition of CNS 11643-15 charset
* lisp/international/mule-conf.el (chinese-cns11643-15): Fix
:code-offset value. (Bug#60275)
* lisp/international/characters.el: Add chinese-cns11643-15 to
charsets whose characters have categories c and C.
* lisp/progmodes/c-ts-mode.el (c-ts-mode--indent-styles): Use new
matcher and anchor.
(c-ts-mode--looking-at-star): New matcher.
(c-ts-mode--comment-start-after-first-star): New anchor.
Yuan Fu [Sat, 24 Dec 2022 01:17:25 +0000 (17:17 -0800)]
; Add treesit_recursion_limit
* src/treesit.c (treesit_recursion_limit): New constant.
(treesit_cursor_helper)
(Ftreesit_search_subtree)
(Ftreesit_induce_sparse_tree): Use the new constant.
Yuan Fu [Sat, 24 Dec 2022 01:12:32 +0000 (17:12 -0800)]
Fix block comment indent and filling for c-ts-mode (bug#59763)
Now indent and filling works like in c-mode. The only noticeable
missing piece is that the "*/" is not attached to the last sentence
when filling. c-mode does it by replacing whitespaces between the
"*/" and the end of the last sentence with xxx, fill it, then change
the xxx back. I don't know if we should do that in c-ts-mode's filling.
* doc/lispref/modes.texi (Parser-based Indentation): Add new preset.
* lisp/progmodes/c-ts-mode.el (c-ts-mode--indent-styles): Add new
indent rule.
(c-ts-mode--fill-paragraph): New function.
(c-ts-base-mode): Setup paragraph-start, adaptive-fill, etc.
* lisp/treesit.el (treesit-simple-indent-presets): Add new preset.
Yuan Fu [Fri, 23 Dec 2022 23:22:31 +0000 (15:22 -0800)]
Fix treesit_cursor_helper (bug#60267)
The cause of that bug is that in a particular parse tree, the node
treesit_cursor_helper tries to go to is a missing node, not only is it
a missing node, it is the first node of a subtree. So when
treesit_cursor_helper follows the algorithm and goes down the tree, it
goes down the previous subtree (because that subtree's end = end_pos,
because the target node has zero width).
o
|
o--+-o
| |
+-+ +-+-+
| | | | |
o x t o o
(We ended up in x when the target is t, because t has zero width.)
One way to solve it is to go back up the tree if we are at a leaf node
and still haven't matched the target node. That's too ugly and
finicky so I resorted to recursion. Now one more functions will
return give up (treesit_node_parent) if we are in a werid parse tree
that is super deep. But since we already kind of give up on this kind
of parse trees (bug#59426), it doesn't really hurt.
* src/treesit.c (treesit_cursor_helper_1): New function.
(treesit_cursor_helper): Use the new function. Change return type to
bool, and accept a cursor pointer.
(Ftreesit_node_parent)
(Ftreesit_search_subtree)
(Ftreesit_search_forward)
(Ftreesit_induce_sparse_tree): Use the new signature.
glacials [Thu, 22 Dec 2022 20:09:08 +0000 (12:09 -0800)]
Correct wrong info in (info)Go to node
The node (info)Go to node in the Info manual states that 'g'
does not allow the use of abbreviations, however it does.
To test this, type 'gt' from this node and see that it takes
you to (info)Top, then type 'ggo<RET>' and see that it takes
you back to (info)Go to node. Tested on emacs 28.2.
* doc/misc/info.texi (Go to node): Fix inaccurate information.
(Bug#60263)
Daniel Martín [Thu, 22 Dec 2022 18:10:24 +0000 (19:10 +0100)]
Add some diff-fixup-modifs tests
* test/lisp/vc/diff-mode-tests.el (diff-mode-test-fixups-added-lines):
Test that diff-mode fixes patches with added lines correctly.
* test/lisp/vc/diff-mode-tests.el (diff-mode-test-fixups-empty-hunks):
Ditto for patches with empty hunks. (Bug#60259)
Ulrich Müller [Mon, 19 Dec 2022 15:51:20 +0000 (16:51 +0100)]
Fix quoted argument in emacsclient-mail.desktop Exec key
Apparently the emacsclient-mail.desktop file doesn't conform to the
Desktop Entry Specification at
https://specifications.freedesktop.org/desktop-entry-spec/desktop-entry-spec-latest.html#exec-variables
which says about the Exec key:
| Field codes must not be used inside a quoted argument, the result of
| field code expansion inside a quoted argument is undefined.
However, the %u field code is used inside a quoted argument of the
Exec key in both the [Desktop Entry] and [Desktop Action new-window]
sections.
* etc/emacsclient-mail.desktop (Exec): The Desktop Entry
Specification does not allow field codes like %u inside a quoted
argument. Work around it by passing %u as first parameter ($1)
to the shell wrapper.
* etc/emacsclient.desktop (Exec): Use `sh` rather than `placeholder`
as the command name of the shell wrapper. (Bug#60204)
Richard Hansen [Sat, 17 Dec 2022 23:51:33 +0000 (18:51 -0500)]
ert-x: Move window selection logic to its own macro
* lisp/emacs-lisp/ert-x.el (ert-with-buffer-selected): New macro to
temporarily display a buffer in a selected window and evaluate a body.
(ert-with-test-buffer-selected): Use the new macro.
* test/lisp/whitespace-tests.el
(ert-test-with-buffer-selected/current)
(ert-test-with-buffer-selected/selected)
(ert-test-with-buffer-selected/nil-buffer)
(ert-test-with-buffer-selected/modification-hooks)
(ert-test-with-buffer-selected/read-only)
(ert-test-with-buffer-selected/return-value): Add tests.
(Bug#60189)
Richard Hansen [Sat, 17 Dec 2022 23:26:33 +0000 (18:26 -0500)]
; ert-x: Add test for buffer read-only state
This test should have been included with commit 29b7d740006fe2190a729bd1c30ccab9356cee36.
* test/lisp/emacs-lisp/ert-x-tests.el
(ert-test-with-test-buffer-selected/read-only): New test.
(Bug#60189)
Eli Zaretskii [Fri, 23 Dec 2022 14:41:08 +0000 (16:41 +0200)]
Fix "C-h k" in recursive minibuffers
* lisp/subr.el (event--posn-at-point): Leave POSN alone if it
doesn't have at least 6 members. This follows more faithfully
what 'event-start' and 'event-end' did before they started using
this function, see commit c1cead89f5f. Call posn-at-point with
the minibuffer-window when in the minibuffer. (Bug#60252)
Michael Albinus [Thu, 22 Dec 2022 15:46:00 +0000 (16:46 +0100)]
Update Tramp version (don't merge with master)
* doc/misc/trampver.texi:
* lisp/net/trampver.el: Change version to "2.6.0.29.1".
(customize-package-emacs-version-alist):
Adapt Tramp version integrated in Emacs 29.1.
João Távora [Thu, 22 Dec 2022 11:29:49 +0000 (11:29 +0000)]
Prevent stale servers when using eglot-extend-to-xref
A weak-valued hash-table is not enough to guarantee that a reference
to a zombie server in eglot--servers-by-xrefed-file variable won't
survive long enough to confuse the next call to eglot--current-server
in some buffers.
So, before this fix it was common to get "Process EGLOT ... not
running" errors if some xref-extended buffers (like system libraries)
were open and M-x eglot-reconnect was issued. This should be
prevented now.
Note however, that even after this the eglot-extend-to-xref logic is
still flawed. For example, if a buffer for the xref-extended buffer
happens to be already visited by the time M-. is issued to navigate to
it, Eglot won't be activated. A half-decent workaround is to kill the
buffer and re-visit it.
* lisp/progmodes/eglot.el (eglot--servers-by-xrefed-file): Move
up.
(eglot--on-shutdown): Make sure to cleanup
eglot--servers-by-xrefed-file.
Jim Porter [Wed, 21 Dec 2022 00:20:50 +0000 (16:20 -0800)]
When redirecting in Eshell, check for "/dev/null" specifically
This is so that users can type "cmd ... > /dev/null" in Eshell no
matter what their system's null device is called. (Users can still
use their system's null device name when redirecting, too. Eshell
doesn't need to do anything special to support that.) This partially
reverts 67a8bdb90c9b5865b7f17290c7135b1a5458c36d. See bug#59545.
Do not merge to master.
* lisp/eshell/esh-io.el (eshell-set-output-handle): Use "/dev/null"
literally.
Eli Zaretskii [Wed, 21 Dec 2022 19:23:51 +0000 (21:23 +0200)]
Fix write-region to null device on MS-Windows
* src/fileio.c (write_region) [WINDOWSNT]: Ignore EBADF errors
from fsync -- this means fsync is not supported for this file.
Happens, for example, with the null device. (Bug#59545)
Richard Hansen [Tue, 13 Dec 2022 06:33:43 +0000 (01:33 -0500)]
whitespace: Avoid mutating original buffer's markers in clones
* lisp/whitespace.el (whitespace--clone): New hook function that is
run after cloning a buffer that copies `whitespace-bob-marker' and
`whitespace-eob-marker' and changes the copies to point to the new
buffer (Bug#59618).
(whitespace-color-on): Register the hook function.
(whitespace-color-off): Unregister the hook function.
* test/lisp/whitespace-tests.el
(whitespace-tests--with-test-buffer): New macro.
(whitespace-tests--check-markers): New function.
(whitespace-tests--indirect-clone-breaks-base-markers)
(whitespace-tests--indirect-clone-markers)
(whitespace-tests--regular-clone-markers): New tests.
Charl P. Botha [Sat, 10 Dec 2022 17:09:38 +0000 (19:09 +0200)]
Fix empty pairs in js tree-sitter imenu alist (bug#59945)
The current js--treesit-imenu, used by the JavaScript, TypeScript and
TSX tree-sitter modes, would return empty pairs in the imenu alist if
there were none of that type of symbol.
This would break both the built in imenu and also packages like
consult-imenu.
See https://github.com/minad/consult/issues/697 for the discussion
there.
Juri Linkov [Tue, 20 Dec 2022 17:22:15 +0000 (19:22 +0200)]
* lisp/repeat.el: Fix repeat-keep-prefix to allow customizing it to non-nil.
* lisp/repeat.el (repeat-keep-prefix): Add or remove
'repeat-pre-hook' depending on the customized value.
(repeat-mode): Add or remove 'repeat-pre-hook' to/from
'pre-command-hook' when 'repeat-keep-prefix' is non-nil.
(repeat-pre-hook): New function.
(repeat-get-map, repeat-check-map): New function refactored from
'repeat-post-hook'.
(repeat-post-hook): Move some code to smaller functions.
(describe-repeat-maps): Set outline-regexp without ^L.
* test/lisp/repeat-tests.el (repeat-tests-keep-prefix):
Uncomment test case that is fixed now in bug#51281 and bug#55986.
Dmitry Gutov [Mon, 19 Dec 2022 19:01:27 +0000 (21:01 +0200)]
ruby-method-params-indent: New user option
* lisp/progmodes/ruby-mode.el (ruby-method-params-indent):
New option (bug#60110).
(ruby-smie-rules): Use it.
* etc/NEWS: Mention it.
* test/lisp/progmodes/ruby-mode-resources/ruby.rb:
Ensure the var's value is default.
* test/lisp/progmodes/ruby-mode-resources/ruby-method-params-indent.rb:
New file.
* test/lisp/progmodes/ruby-mode-tests.el (ruby-deftest-indent):
New macro, use it to run the indentation test using the new file.
Disable the :expensive-test tag, because neither runs for "longer
than some few seconds", both take significantly below 1s.
Eli Zaretskii [Mon, 19 Dec 2022 17:01:04 +0000 (19:01 +0200)]
Fix storing email into nnmail by Gnus
* lisp/gnus/nnml.el (nnml--encode-headers): Wrap
'rfc2047-encode-string' calls with 'ignore-errors', to avoid
disrupting email workflows due to possibly-invalid headers.
Reported by Florian Weimer <fweimer@redhat.com>.
F. Jason Park [Mon, 12 Dec 2022 03:16:07 +0000 (19:16 -0800)]
; Really respect browse-url var in erc-compat
* lisp/erc/erc-compat.el: Do what was supposed to be done by 75f26646d4a569cfb485de4baddcda66ff44b2c3 "; Be nicer when updating
browse-url var in erc-compat". This is the less harmful version of
that patch (from bug#59976#8) but without the cl-lib requirement since
users may not want to load the main ERC library right away.
* lisp/erc/erc.el: Clarify some comments regarding the core API.
Richard Hansen [Mon, 19 Dec 2022 04:04:00 +0000 (23:04 -0500)]
whitespace: Fix unintended change in buffer modification status
* lisp/whitespace.el (whitespace--empty-at-bob-matcher)
whitespace--empty-at-eob-matcher, whitespace--update-bob-eob):
Silently add the `font-lock-multiline' text property when
highlighting beginning-of-buffer and end-of-buffer empty lines
to prevent Emacs from running modification hooks or considering
the buffer to be modified (Bug#60066).
* test/lisp/whitespace-tests.el
(whitespace-tests--empty-bob-eob-modified): Add a regression test.
Repair setopt test after error demotion to warning
* test/lisp/cus-edit-tests.el (test-setopt):
Check for a warrning instead of an error in attempt to call `setopt`
with a value that does not match the declared type (bug#60162).
Dmitry Gutov [Sun, 18 Dec 2022 12:01:47 +0000 (14:01 +0200)]
ruby-mode: Support endless methods (bug#54702)
* lisp/progmodes/ruby-mode.el (ruby-endless-method-head-re):
New constant.
(ruby-smie-grammar): New token.
(ruby-smie--forward-token, ruby-smie--backward-token):
Recognize it.
(ruby-smie-rules): Indentation support.
(ruby-add-log-current-method): Support here too.
* test/lisp/progmodes/ruby-mode-tests.el
(ruby-add-log-current-method-after-endless-method): New test.
* test/lisp/progmodes/ruby-mode-resources/ruby.rb: New examples.
Gregory Heytings [Sun, 18 Dec 2022 10:46:48 +0000 (10:46 +0000)]
Improve and extend admin/git-bisect-start
* admin/git-bisect-start: Use 'git bisect skip' instead of 'git
bisect good' for merges of external trees, which is equivalent but
faster in some cases. Improve the explanation on external tree
merges. Add list of commits in the last six years on which
building Emacs fails, and call 'git bisect skip' on them. Also
document how the script can be called automatically when 'git
bisect start' is called in the Emacs repository.
* admin/notes/repo (Bisecting): Adapt the explanation of the
script accordingly.
Andrea Monaco [Fri, 9 Dec 2022 20:22:22 +0000 (21:22 +0100)]
Make 'rmail-summary-by-thread' faster
* lisp/mail/rmailsum.el (rmail-summary-message-parents-vector)
(rmail-summary-message-descendants-vector): Doc fixes.
(rmail-summary-message-descendants-vector): New variable.
(rmail-summary-fill-message-parents-and-descs-vectors): Renamed
from 'rmail-summary-fill-message-parents-vector' and rewritten.
(rmail-summary-direct-descendants): Function deleted.