Grudgingly accept function values in the function position
* lisp/emacs-lisp/cconv.el (cconv-convert):
Warn about (F ...) where F is a non-symbol function value (bytecode
object etc), but let it pass for compatibility's sake (bug#68931).
* test/lisp/emacs-lisp/bytecomp-tests.el (bytecomp--fun-value-as-head):
New test.
Po Lu [Mon, 5 Feb 2024 10:34:22 +0000 (18:34 +0800)]
Fix frame focus tracking under Android
* java/org/gnu/emacs/EmacsActivity.java (invalidateFocus): New
argument WHENCE, a unique number identifying the circumstances
leading up to the call. All callers changed.
(attachWindow): Call `invalidateFocus' from the UI thread.
(onWindowFocusChanged): Don't remove activity from
`focusedActivities' if it already exists should `hasWindowFocus'
return true.
Yuan Fu [Sun, 4 Feb 2024 05:24:29 +0000 (21:24 -0800)]
Improve treesit-forward-sexp behavior for leaf nodes (bug#68899)
treesit-forward-sexp uses treesit--navigate-thing with 'restricted'
tactic. In this tactic we don't move over the parent thing. However,
this makes forward-sexp useless for symbols when point is in the
symbol rather than at the beginning of it: in that case, the symbol is
considered parent and treesit-forward-sexp won't move to the end of
it.
To solve that, we allow to move across the parent even in 'restricted'
mode if the parent is a leaf thing.
Here, "leaf thing" is defined as "doesn't have any child 'thing'
inside it".
* lisp/treesit.el (treesit--navigate-thing): Move over parent in
'restricted' tactic if the parent is a leaf thing.
The assertion failure was raised at lread.c:411 during the
`lread-invalid-bytecodes` test in `test/src/lread-tests.el`.
I suspect we could remove the assertion instead.
Stefan Monnier [Sun, 4 Feb 2024 17:50:55 +0000 (12:50 -0500)]
(w->base_line_number): Rework the way we flush the cache
* src/xdisp.c (BASE_LINE_NUMBER_VALID_P): New macro.
(try_scrolling): Use it.
(redisplay_window, Fformat_mode_line): Use it to flush the
base_line_number (if it's stale) once at the beginning.
(decode_mode_spec): Don't use (or set) `w->start` and
`w->base_line_number` when operating on another buffer!
Eshel Yaron [Tue, 6 Feb 2024 19:33:53 +0000 (20:33 +0100)]
Support bookmarking Xref results buffers
* lisp/progmodes/xref.el (bookmark-make-record-default)
(bookmark-make-record, bookmark-prop-get)
(bookmark-handle-bookmark, bookmark-get-rear-context-string)
(bookmark-get-front-context-string): Declare functions.
(xref-backend-context, xref-backend-restore): New generic functions.
(xref--backend, xref--identifier, xref--kind)
(xref--original-buffer, xref--original-point): New local variables.
(xref--show-common-initialize): Set them in Xref results buffer.
(xref-default-bookmark-name-format): New user option.
(xref-bookmark-make-record, xref-bookmark-jump): New functions.
(xref--xref-buffer-mode): Set 'bookmark-make-record-function'.
(xref-fetcher-alist): New variable.
(xref--show-xref-buffer, xref-show-definitions-buffer)
(xref-show-definitions-buffer-at-bottom): Use it.
(xref--read-identifier): Improve error message.
(xref-make-fetcher): Extract from...
(xref--create-fetcher): ...here.
* doc/emacs/maintaining.texi (Xref Commands): Document new feature.
* etc/NEWS: Announce new feature and Xref generic functions.
Eshel Yaron [Sun, 4 Feb 2024 12:57:06 +0000 (13:57 +0100)]
; Fix remote shell file name selection
* lisp/shell.el (shell): Cease providing 'file-remote-p' as completion
predicate when reading remote shell file name. The completion predicate
in 'read-file-name' gets the non-directory part of the file name, which
in particular does not include the remote part, if any, so this
predicate always fail in these settings. Moreover, this predicate
didn't make a lot of sense to begin with.
Stefan Kangas [Sun, 4 Feb 2024 09:28:40 +0000 (10:28 +0100)]
Delete compat code in `url` library
* lisp/url/url-cid.el (url-cid): Delete compat code for ancient Gnus.
* lisp/url/url-ldap.el (url-ldap-certificate-formatter): Delete compat
code; ssl.el has never been in Emacs.
* lisp/url/url-mailto.el (url-mail): Make into alias for 'message-mail',
since it is always fboundp.
Eli Zaretskii [Sun, 4 Feb 2024 09:45:15 +0000 (11:45 +0200)]
Fix display of invisible text with opposite directionality
* src/xdisp.c (handle_invisible_prop): Skip invisible text
correctly when it starts at position whose resolved bidi level is
above the base paragraph level. (Bug#68446)
Stefan Monnier [Sun, 4 Feb 2024 04:05:03 +0000 (23:05 -0500)]
tramp: Tweak the ls-lisp declarations
* lisp/net/tramp-sh.el (ls-lisp-use-insert-directory-program): Don't
declare its existence...
(tramp-sh-handle-insert-directory): ...test it instead.
* lisp/net/tramp.el (ls-lisp-dirs-first, ls-lisp-emulation)
(ls-lisp-ignore-case, ls-lisp-use-insert-directory-program)
(ls-lisp-verbosity): Move declaration...
(tramp-handle-insert-directory): ...to the point where we have a good
reason to think these variables exist.
João Távora [Thu, 30 Nov 2023 13:32:50 +0000 (07:32 -0600)]
Fix prefix discovery for files with read-symbol-shorthands (bug#67325)
In a previous commit, the local-variable read-symbol-shorthands is
already read into the temporary buffer used for the autoload parsing
aerobatics, so all we needed to do in 'l-g--compute-prefixes' is
use 'read' to give 'read-symbol-shorthands' a chance to kick in.
Add font locking to the shorthand prefix of a given printed symbol
name by checking if any of the shorthand prefixes in
read-symbol-shorthands are a prefix for that print name. Although
this does more string comparisons, it didn't prove to be any slower
than the existing approach, and is more correct.
This version is more accurate when highlighting files with many
overlapping shorthands. Given:
Eli Zaretskii [Sat, 3 Feb 2024 11:00:15 +0000 (13:00 +0200)]
Fix downcasing of mode-name in compile.el
* lisp/progmodes/compile.el (compilation--downcase-mode-name): New
function.
(compilation-start, kill-compilation): Use it instead of calling
'downcase' on 'mode-name'. (Bug#68795)
Fix incompatibility with tree-sitter-javascript >= 0.20.2
Starting from version 0.20.2 the grammar's primary expression
"function" has been renamed to "function_expression". A new
function checks if the new primary expression is available,
and if so, it returns the correct rules.
* lisp/progmodes/js.el
(js--treesit-font-lock-compatibility-definition-feature): New
function.
(js--treesit-font-lock-settings): Use it. (Bug#68879)
Eli Zaretskii [Sat, 3 Feb 2024 09:09:36 +0000 (11:09 +0200)]
Avoid signaling errors from 'pixel-fill-region'
* lisp/textmodes/pixel-fill.el (pixel-fill-region): Make sure the
selected window displays the current buffer. This is important
when this function is called inside 'with-current-buffer' or
similar forms which temporarily change the buffer displayed in the
selected window. (Bug#67791)
F. Jason Park [Wed, 31 Jan 2024 14:01:54 +0000 (06:01 -0800)]
Reassociate erc-networks--id for orphaned queries
* lisp/erc/erc-networks.el (erc-networks--examine-targets): Adopt the
server's network ID in query buffers created before MOTD's end. Do
this to avoid a type error in the process filter when renaming
buffers.
* lisp/erc/erc-networks.el (erc-networks--examine-targets): New test.
* test/lisp/erc/erc-scenarios-base-upstream-recon-znc.el
(erc-scenarios-upstream-recon--znc/severed): New test.
* test/lisp/erc/erc-scenarios-misc.el
(erc-scenarios-base-mask-target-routing): Adjust timeout.
* test/lisp/erc/resources/base/upstream-reconnect/znc-severed.eld:
New file.
* test/lisp/erc/resources/erc-tests-common.el
(erc-tests-common-make-server-buf): Use NAME parameter for creating
ID.
F. Jason Park [Wed, 31 Jan 2024 02:17:41 +0000 (18:17 -0800)]
Fix local variable persistence in erc-stamp
* etc/ERC-NEWS: Mention renaming of `erc-munge-invisible-property'.
* lisp/erc/erc-stamp.el (erc-stamp-mode, erc-stamp-disable): Remove
correct function from `erc-mode-hook'.
(erc-stamp--recover-on-reconnect): Revise doc string.
(erc-munge-invisibility-spec, erc-stamp--manage-local-options-state):
Mark former name as obsolete and rename to latter. Don't use helper
macro meant only for local modules. This bug originated from c68dc778
"Manage some text props for ERC insertion-hook members", which stemmed
from bug#60936.
(erc-stamp--setup, erc-hide-timestamps, erc-show-timestamps)
(erc-toggle-timestamps): Use new name for
`erc-munge-invisibility-spec'.
* lisp/erc/erc.el (erc--restore-initialize-priors): Raise error at
runtime if mode var doesn't belong to a local module.
* test/lisp/erc/erc-stamp-tests.el (erc-stamp-tests--insert-right)
(erc-timestamp-intangible--left): Use new name for
`erc-munge-invisibility-spec'.
* test/lisp/erc/erc-tests.el (erc--refresh-prompt): Shadow
`erc-last-input-time'.
(erc--restore-initialize-priors): Add error form to expected
expansion, and skip test on Emacs 27.
* test/lisp/erc/resources/erc-scenarios-common.el
(erc-scenarios-common--make-bindings): Shadow `erc-last-input-time'.
F. Jason Park [Tue, 30 Jan 2024 03:04:58 +0000 (19:04 -0800)]
Teach customize-option about erc-modules
* lisp/erc/erc-goodies.el (erc-scrolltobottom-mode)
(erc-scrolltobottom-enable): Use `setq' instead of `setopt' because
the latter isn't defined in Emacs 27 and 28. This fix is unrelated to
the main thrust of this commit.
* lisp/erc/erc.el (erc-modules): Make good on decades old language in
info node "(erc) Modules" by ensuring `customize-option' can find this
option before its containing library is loaded. Like
`gnus-select-method', this option serves as an entry point for
configuring the application and is presented that way in tutorials and
library front matter. Moreover, it can't be reasonably autoloaded in
the traditional way because of its many dependencies and large textual
footprint.
(erc-display-message): Revise doc string.
Stefan Monnier [Fri, 2 Feb 2024 23:59:21 +0000 (18:59 -0500)]
cperl-mode.el: Don't use obsolete `special-display-popup-frame`
* lisp/progmodes/cperl-mode.el (cperl-info-on-command): Simplify,
to let `pop-to-buffer` decide whether to create a new frame or not,
so it can be controlled by `display-buffer-alist`.
Stefan Monnier [Wed, 31 Jan 2024 23:56:43 +0000 (18:56 -0500)]
bytecomp.el: Rewrite the way we print dynamic docstrings
We used to print dynamic docstrings "manually" for two reasons:
- References should look like `(#$ . POS)` but `prin1` was unable
to print just `#$` for an sexp.
- `make-docfile` needed to find those docstrings and the object
to which they belonged.
The second point is moot now that we don't use `make-docfile` on
`.elc` files. So this patch lifts the first restriction,
using `print-number-table`.
The rest of the patch then simplifies and regularises the
bytecompiler's generation of dynamic docstrings, which can
now also easily be done for "inner" defvars and other places.
* lisp/emacs-lisp/bytecomp.el:
(byte-compile--list-with-n): New function.
(byte-compile--docstring-style-warn): Rename from
`byte-compile-docstring-style-warn` and change calling convention.
(byte-compile--\#$, byte-compile--docstrings): New vars.
(byte-compile-close-variables): Bind them.
(byte-compile--docstring): New function.
(byte-compile-from-buffer): Set `byte-compile--\#$`.
(byte-compile-output-file-form): Use `byte-compile--\#$` instead
of special casing specific forms.
(byte-compile--output-docform-recurse, byte-compile-output-docform):
Delete functions.
(byte-compile-file-form-autoload, byte-compile-file-form-defalias)
(byte-compile-file-form-defvar-function, byte-compile-lambda):
Use `byte-compile--docstring` and `byte-compile--list-with-n`.
(byte-compile--declare-var): Add optional `not-toplevel` arg.
(byte-compile-defvar): Add `toplevel` arg. Use `byte-compile--docstring`.
(byte-compile-file-form-defvar): Delegate to `byte-compile-defvar`.
(byte-compile--custom-declare-face): New function. Use it for
`custom-declare-face`.
(byte-compile-file-form-defmumble): Use `byte-compile-output-file-form`
* src/doc.c (Fdocumentation_stringp): New function.
(syms_of_doc): Defsubr it.
(store_function_docstring): Remove left-over code from when we
used DOC for the docstring of some Lisp files.
Stefan Kangas [Fri, 2 Feb 2024 11:28:54 +0000 (12:28 +0100)]
Increase `emacs-lisp-docstring-fill-column` to 72
Monitors are wider now than when these defaults were first set, and it
is useful to take better advantage of that, to fit text on fewer lines.
Yet, it has repeatedly been shown that overly long lines reduce
readability:
"A reasonable guideline would be 55 to 75 characters per line."[1]
We also don't want to disfavor narrow displays, like mobile phones; a
more promising direction here might be to automatically word wrap
docstrings and make their maximum width customizable. That might
require a new docstring format, however.
Bumping it by 7 characters, from 65 to 72, seems a reasonable compromise
for now. Consideration was given to increasing it to 70 or 75, but 72
happens to be a commonly recommended maximum line width elsewhere (see
Fortran 66, Python docstrings, commit message recommendations, etc.),
and we might as well do the same.
This change was discussed in:
https://lists.gnu.org/r/emacs-devel/2022-07/msg00217.html
[1] "Optimal Line Length in Reading — A Literature Review", Nanavati and
Bias, Visible Language, Vol. 39 No. 2 (2005).
https://journals.uc.edu/index.php/vl/article/view/5765
Sacha Chua [Fri, 26 Jan 2024 13:54:03 +0000 (08:54 -0500)]
shr: Correct SVG attribute case
* lisp/net/shr.el (shr-correct-attribute-case): New constant.
(shr-correct-dom-case): New function to correct SVG attribute case.
(shr-tag-svg): Correct SVG attribute cases before using them.
Eshel Yaron [Sat, 3 Feb 2024 15:06:02 +0000 (16:06 +0100)]
Support sorting bookmark completions by last modified
* lisp/bookmark.el (bookmark-sort-by-last-modified-time): New fun.
(bookmark-completing-read): Provide it as a sorting option via
'minibuffer-completions-sort-orders'.
Po Lu [Thu, 1 Feb 2024 08:16:09 +0000 (16:16 +0800)]
Prevent continuation from affecting tab width in/after line prefix
* src/dispextern.h (struct it) <wrap_prefix_width>: New field,
synchronized with current_x when producing glyphs for wrap
prefixes, and subtracted from it->current_x when computing tab
widths.
* src/term.c (produce_glyphs): Set wrap_prefix_width.
* src/xdisp.c (start_display, display_min_width, move_it_to)
(move_it_vertically_backward, move_it_by_lines)
(window_text_pixel_size, display_tab_bar_line)
(display_tool_bar_line, redisplay_internal, redisplay_window)
(try_window_id, insert_left_trunc_glyphs)
(extend_face_to_end_of_line, display_line)
(Fmove_point_visually): Set or clear wrap_prefix_width as
appropriate.
(gui_produce_glyphs): Set or clear it->wrap_prefix_width. When
computing the base position of a tab character, do not subtract
the continuation line width if a line prefix is the current
iterator method. Subtract the wrap_prefix_width otherwise, in
order that the width of the tab is computed free of influence
from the wrap prefix.
Eshel Yaron [Fri, 2 Feb 2024 12:33:43 +0000 (13:33 +0100)]
; Fix updating completion preview after deletion at eob
* lisp/completion-preview.el (completion-preview--show): Use the
actual overlay position, instead of its 'completion-preview-end'
property, in case the overlay has moved (e.g. when called after a
deletion command).
This fixes an issue where 'completion-preview-next-candidate' would
fail to take into account the part of the symbol that follows
point (the suffix) when point is at the middle of a symbol, as well as
a similar issue in 'completion-preview--show' that would manifest with
slow 'completion-at-point-functions'.
* lisp/completion-preview.el (completion-preview-next-candidate)
(completion-preview--show): Use recorded 'completion-preview-end'
position instead of current point.
* test/lisp/completion-preview-tests.el (completion-preview-mid-symbol-cycle):
New test.
Bytecode engine fast-path streamlining of plain symbols
* src/bytecode.c (exec_byte_code):
Only use fast-path optimisations for calls and dynamic variable
reference and setting where the symbol is plain, which is much faster.
The obsolete lazy-loaded bytecode feature, enabled by
`byte-compile-dynamic`, slows down Lisp execution even when not in use
because every call to a bytecode function has to check that function
for laziness.
This change forces up-front loading of all lazy bytecode so that we
can remove all those checks. (Dynamically loaded doc strings are not
affected.)
There is no point in generating lazy bytecode any more so we stop
doing that; this simplifies the compiler. `byte-compile-dynamic` now
has no effect.
This is a fully compatible change; the few remaining users of
`byte-compile-dynamic` should not notice any difference.
* src/lread.c (bytecode_from_rev_list): Force eager loading of
lazy bytecode.
* src/bytecode.c (exec_byte_code): Remove lazy bytecode checks.
* src/eval.c (fetch_and_exec_byte_code, Ffetch_bytecode): Remove.
(funcall_lambda): Call exec_byte_code directly, avoiding checks.
* lisp/subr.el (fetch-bytecode): New definition, obsolete no-op.
* lisp/emacs-lisp/disass.el (disassemble-1):
* lisp/emacs-lisp/bytecomp.el (byte-compile-unfold-bcf):
Remove calls to fetch-bytecode.
(byte-compile-dynamic): Update doc string.
(byte-compile-close-variables, byte-compile-from-buffer)
(byte-compile-insert-header, byte-compile-output-file-form)
(byte-compile--output-docform-recurse, byte-compile-output-docform)
(byte-compile-file-form-defmumble):
Remove effects of byte-compile-dynamic.
* doc/lispref/compile.texi (Dynamic Loading): Remove node now that
the entire `byte-compile-dynamic` facility has been rendered inert.
* etc/NEWS: Announce changes.
Allow equal user-defined hash table tests with different names
Hash tables using different user-defined tests defined identically
sometimes ended up using the wrong test (bug#68668).
* src/fns.c (get_hash_table_user_test): Take test name into account
when matching the test object.
* test/src/fns-tests.el (fns--define-hash-table-test): New.
Po Lu [Wed, 31 Jan 2024 06:34:19 +0000 (14:34 +0800)]
Don't hang when display objects are displaced by line or wrap-prefixes
This fixes a hang that would frequently rear its ugly head while
displaying messages in the `telega.el' instant messenger client,
which inserts images approaching the width of the window with
line and wrap prefixes.
* src/xdisp.c (move_it_in_display_line_to): If a line or wrap
prefix is set in place, do not generate continuation lines until
a minimum of one glyph has been produced outside that prefix.
(move_it_to): Remove the previous workaround that could not
recover from errors caused by display strings.
(display_line): Synchronize with move_it_in_display_line_to;
remove old workaround that only provided for oversized wrap
prefixes comprising `space' display objects.
Suppose the block comment is covered by a local parser. When we
indent line 3, treesit--indent-1 will try to get the local parser at
the BOL, and it'll get the local parser. But it shouldn't use the
local parser to indent this line, it should use the host parser of
that local parser instead.
So now, if treesit--indent-1 gets a local parser, but the local
parser's root node's start coincides with BOL, treesit--indent-1 will
use the host parser to indent this line.
We also need to make treesit--update-ranges-local to save the host
parser along with the local parser, and make
treesit-local-parsers-at/on extract and return the host parser.
I also switch the two cases in the cond form in treesit--indent-1:
(null (treesit-parser-list)) and (car local-parsers), (car
local-parsers) now takes precedence.
* lisp/treesit.el (treesit-local-parsers-at):
(treesit-local-parsers-on): Add WITH-HOST parameter.
(treesit--update-ranges-local): Save the host parser to the local
overlay.
(treesit--indent-1): If the root node of the local parser is at BOL,
use the host parser instead.