Po Lu [Fri, 13 Oct 2023 03:57:46 +0000 (11:57 +0800)]
Try harder to generate tombstones upon emacs_abort
* src/android.c (HAS_BUILTIN_TRAP): New macro definition.
(emacs_abort) [HAS_BUILTIN_TRAP]: Call __builtin_trap to abort,
in place of dereferencing NULL.
* src/androidterm.c (android_draw_fringe_bitmap): Correct
comment transplanted from X code.
Po Lu [Thu, 12 Oct 2023 13:46:54 +0000 (21:46 +0800)]
Enable highlighting Gud execution lines without hl-line-mode
* doc/emacs/building.texi (Debugger Operation): Mention two new
options and the relationship between Gud and HL Line Mode.
* etc/NEWS (Editing Changes in Emacs 30.1): Mention the new
option.
* lisp/progmodes/gud.el (gud-highlight-current-line-overlay):
New variable.
(gud-sentinel) <signal, exit>: Delete that overlay if set.
(gud-highlight-current-line, gud-highlight-current-line-face):
New user options.
(gud-display-line): Create and move an overlay without employing
hl-line-mode, if so enjoined by the user.
* test/lisp/filenotify-tests.el (file-notify-test04-autorevert):
Use `skip-when'.
(file-notify-test12-unmount, file-notify-test12-unmount-remote):
New tests.
Po Lu [Thu, 12 Oct 2023 00:43:56 +0000 (08:43 +0800)]
Respond to JNI errors around drawing operations
* src/android.c (android_fill_polygon, android_draw_rectangle)
(android_draw_point, android_draw_line, android_lock_bitmap):
Check or clear errors around potential JNI errors; the penalty
incurred to performance is not as significant as was expected.
Alan Mackenzie [Wed, 11 Oct 2023 13:26:01 +0000 (13:26 +0000)]
In cl-prin1, enable raw printing for a byte-compiled function
* lisp/emacs-lisp/cl-print.el (cl-print-compiled): document the
new option `raw'.
(cl-print-object/compiled-function): when cl-print-compiled is
`raw', just print the function using `prin1'. Apply a button
to this output which, when activated disassembles the function.
* etc/NEWS (cl-print): Add an entry for this new feature.
These were there to work around deficiencies in how Eshell produces
completions for 'pcomplete-argument' (Eshell passed various non-string
objects to Pcomplete, which broke things). Now, Eshell always returns
a stringified form of the argument, with the original value stored via
the text property 'pcomplete-arg-value'.
* lisp/pcomplete.el (pcomplete-arg): Revert changes back to a simpler
form.
Ulrich Müller [Tue, 10 Oct 2023 09:58:03 +0000 (11:58 +0200)]
Update astronomical length units in Calc
* lisp/calc/calc-units.el (math-standard-units): Update the
astronomical unit and the parsec, using their definitions by the
International Astronomical Union.
* lisp/progmodes/project.el (project-mode-line): New user option.
(project-menu-entry, project-mode-line-map): New variables.
(project-mode-line-format): New variable.
(project-mode-line-format): New function.
Juri Linkov [Mon, 9 Oct 2023 18:03:03 +0000 (21:03 +0300)]
* lisp/progmodes/xref.el: Show the number of matches on mode-line (bug#66332)
(xref-num-matches-found, xref-num-matches-face, xref-mode-line-matches):
New variables.
(xref--show-xref-buffer): Set xref-num-matches-found to the length of xrefs,
and buffer-local mode-line-process to xref-mode-line-matches.
Harald Jörg [Mon, 9 Oct 2023 15:49:15 +0000 (17:49 +0200)]
; cperl-mode.el: Don't override fontificaton in comments.
* lisp/progmodes/cperl-mode.el (cperl-init-faces): Reorder the
matchers for fontification of array and hash elements and keys so
that they don't override comment and string fontification (Bug#66145).
* test/lisp/progmodes/cperl-mode-tests.el (cperl-test-bug-66145):
New test for all combinations of sigils and brackets/braces to
verify that strings and comments are left untouched. This test
also works for perl-mode which has always done it correctly.
* test/lisp/progmodes/cperl-mode-resources/cperl-bug-66145.pl: New
resource file for the above test.
Don't allow Services entries in pop-up menus on macOS (bug#66420)
* src/nsmenu.m ([EmacsMenu runMenuAt:forFrame:keymaps:]):
Prevent the system from adding context menu plug-ins
(Services entries or sub-menu) to all our pop-up menus.
* lisp/emacs-lisp/package-vc.el (package-vc--unpack-1): If a file
matches the ignored packages in :ignored-files, do not install any
dependencies it lists.
Po Lu [Sun, 8 Oct 2023 01:13:11 +0000 (09:13 +0800)]
Introduce a tool bar for dictionary mode
* lisp/doc-view.el (doc-view-minor-mode-menu): Correct typo in
edit mode menu.
* lisp/net/dictionary.el (dictionary-mode-menu): New menu.
(dictionary-tool-bar-map): New variable; derive menu bar entries
from the dictionary-mode-menu.
(dictionary-mode): Set the tool bar map to
dictionary-tool-bar-map.
Eli Zaretskii [Sat, 7 Oct 2023 07:48:00 +0000 (03:48 -0400)]
Merge from origin/emacs-29
a74e51cfd15 Fix a defcustom :type c27b90d04bf Fix 'ido--ffap-find-file' 1594d5f17ad Fix setting the pipe capacity for subprocesses aad8b5d78f3 Handle LANG on macOS differently (bug#65908)
Eli Zaretskii [Sat, 7 Oct 2023 06:46:04 +0000 (09:46 +0300)]
Fix 'ido--ffap-find-file'
* lisp/ido.el (ido--ffap-find-file): Make the signature consistent
with that of 'find-file', and pass the optional second argument to
'find-file'. (Bug#66382)
Eli Zaretskii [Fri, 6 Oct 2023 05:31:59 +0000 (08:31 +0300)]
Fix setting the pipe capacity for subprocesses
* src/process.c (create_process) [F_SETPIPE_SZ]: Set the pipe
capacity only if the required read-process-max is larger than the
default capacity of the pipe. (Bug#66288)
Gerd Möllmann [Thu, 5 Oct 2023 18:41:54 +0000 (20:41 +0200)]
Handle LANG on macOS differently (bug#65908)
* src/nsterm.m (ns_init_locale): If LANG is set, try to use that,
otherwise try to deduce what LANG should be. Check is the result is
valid, and use LANG="en_US.UTF-8" if not.
Silence macro expansion during completion at point
* lisp/emacs-lisp/macroexp.el (macroexp-inhibit-compiler-macros): Add
variable.
(macroexp--compiler-macro): Inspect that new variable and, if it is
non-nil, return the input form unchanged.
* lisp/progmodes/elisp-mode.el (elisp--local-variables): Silence
messages. Avoid compiler macros. (Bug#58148)
It keeps interfering with other rules, slowing everything down a
little bit and makes it harder to add or change other rules. The rule
is still there and can easily be re-enabled by those who need it.
* etc/NEWS: Announce.
* lisp/progmodes/compile.el (compilation-error-regexp-alist):
Exclude `omake`.
* test/lisp/progmodes/compile-tests.el
(compile-tests--test-regexps-data):
Actually test the `cucumber` rule. Remove the `omake` test case.
(compile-test-error-regexps):
Test `omake` here. Test other rules without `omake` included.
Spencer Baugh [Wed, 28 Jun 2023 12:48:01 +0000 (08:48 -0400)]
Make newly-created smerge-diff-buffers read-only
Buffers name *vc-diff* are usually created by vc, which makes them
read-only. If we create such a buffer, let's make it read-only too.
If the buffer already exists, though, don't change that since the user
might have deliberately made it writable.
* lisp/vc/smerge-mode.el (smerge-diff): Make newly-created
smerge-diff-buffers read-only. (bug#64071)
* src/androidgui.h (enum android_ime_operation): New operation
ANDROID_IME_REPLACE_TEXT.
* src/androidterm.c (android_handle_ime_event): Decode text when
encountering an ANDROID_IME_REPLACE_TEXT operation. Return if
decoding overflowed rather than presenting Qnil to textconv
functions.
(replaceText): New JNI function.
* src/frame.h (enum text_conversion_operation): New operation
TEXTCONV_REPLACE_TEXT.
* src/textconv.c (really_commit_text): Move point to start if
the composing region is set.
(really_replace_text): New function.
(handle_pending_conversion_events_1) <TEXTCONV_REPLACE_TEXT>:
New case.
(replace_text): New function.
Po Lu [Wed, 4 Oct 2023 08:33:05 +0000 (16:33 +0800)]
Correct local reference leaks
* src/android.c (android_build_string): Accept a list of local
references to destroy upon an allocation failure, facilitating
the proper deallocation of local references in such situations.
(android_browse_url): Revise for new calling convention.
* src/androidmenu.c (android_menu_show, android_dialog_show):
Revise for new calling convention.
* src/androidselect.c (android_notifications_notify_1): Supply
each successive local reference to android_build_string as
notification text is being encoded.
Alan Mackenzie [Tue, 3 Oct 2023 10:22:26 +0000 (10:22 +0000)]
Correct the `cond' forms in cl-print-string-with-limit
In this function, calling with limit bound to t will cause an
error in any of the cond forms which set print-length, etc.
Correct them!
* lisp/emacs-lisp/cl-print.el (cl-print-string-with-limit):
Amend the doc string. In the cond forms in the bindings for
print-length, etc., test the value t first. Amend those for
print-length and print-level also to test for a zero value of
limit.
Jim Porter [Sat, 23 Sep 2023 18:36:11 +0000 (11:36 -0700)]
Support Eshell iterative evaluation in the background
This really just generalizes Eshell's previous support for iterative
evaluation of a single current command to a list of multiple commands,
of which at most one can be in the foreground (bug#66066).
* lisp/eshell/esh-cmd.el (eshell-last-async-procs)
(eshell-current-command): Make obsolete in favor of...
(eshell-foreground-command): ... this
(eshell-background-commands): New variable.
(eshell-interactive-process-p): Make obsolete.
(eshell-head-process, eshell-tail-process): Use
'eshell-foreground-command'.
(eshell-cmd-initialize): Initialize new variables.
(eshell-add-command, eshell-remove-command)
(eshell-commands-for-process): New functions.
(eshell-parse-command): Make 'eshell-do-subjob' the outermost call.
(eshell-do-subjob): Call 'eshell-resume-eval' to split this command
off from its parent forms.
(eshell-eval-command): Use 'eshell-add-command'.
(eshell-resume-command): Use 'eshell-commands-for-process'.
(eshell-resume-eval): Take a COMMAND argument. Return
':eshell-background' form for deferred background commands.
(eshell-do-eval): Remove check for 'eshell-current-subjob-p'. This is
handled differently now.
* lisp/eshell/eshell.el (eshell-command): Wait for all processes to
exit when running synchronously.
* test/lisp/eshell/esh-cmd-tests.el
(esh-cmd-test/background/simple-command)
(esh-cmd-test/background/subcommand): New tests.
(esh-cmd-test/throw): Use 'eshell-foreground-command'.
* test/lisp/eshell/eshell-tests.el (eshell-test/queue-input): Use
'eshell-foreground-command'.
* test/lisp/eshell/em-script-tests.el
(em-script-test/source-script/background): Make the test script more
complex.
* test/lisp/eshell/eshell-tests.el
(eshell-test/eshell-command/pipeline-wait): New test.
* doc/misc/eshell.texi (Bugs and ideas): Remove implemented feature.
Po Lu [Tue, 3 Oct 2023 00:55:57 +0000 (08:55 +0800)]
Update Android port
* src/androidvfs.c (android_afs_open, android_asset_fstat):
Return proper mtime within file descriptors incorporating asset
data and within android_asset_fstat.
* src/xterm.c (handle_one_xevent)
<KeyPress, XI_KeyPress>: Verify presence of a quit keysym
prior to registering it.
F. Jason Park [Sat, 10 Jun 2023 04:00:03 +0000 (21:00 -0700)]
Inhibit slash commands in erc--input-review-functions
* lisp/erc/erc.el (erc--input-review-functions): Add new review
function `erc--inhibit-slash-cmd-insertion'.
(erc--check-prompt-input-functions): Move
`erc--check-prompt-input-for-multiline-command' above
`erc--check-prompt-input-for-multiline-blanks'.
(erc--inhibit-slash-cmd-insertion): New "review" function to suppress
insertion of prompt input for slash commands. Doesn't affect "meta"
slash commands like /SAY.
(erc--send-input-lines): Don't bother checking whether message is a
command. Instead, trust verdict handed down by message-prep and
review functions. This opens the door to optional insertion for
debugging purposes or when echoing command lines in a shell-like
fashion.
* test/lisp/erc/erc-tests.el (erc-send-whitespace-lines): clean up
portion dealing with trimming slash commands. (Bug#66073)
F. Jason Park [Fri, 15 Sep 2023 13:08:55 +0000 (06:08 -0700)]
Improve erc-warn-about-blank-lines behavior
* etc/ERC-NEWS: Mention more detailed feedback when option
`erc-warn-about-blank-lines' is non-nil.
* lisp/erc/erc-common.el (erc--input-split): Add `abortp' slot to
allow a premature exit while validating prompt input.
* lisp/erc/erc.el (erc-warn-about-blank-lines): Clarify meaning of
"blank lines" in doc string, and mention interaction with
companion option `erc-send-whitespace-lines'.
(erc-inhibit-multiline-input): Fix inaccurate description in doc
string.
(erc--input-review-functions): Move
`erc--discard-trailing-multiline-nulls' to end of list, after
`erc--run-input-validation-checks' so that the latter doesn't have to
resplit the original input string in order to issue feedback.
(erc--blank-in-multiline-input-p): Remove unused internal function
originally slated to be part of ERC 5.6.
(erc--check-prompt-input-for-something): New trivial validation
function to check if the input is empty.
(erc--count-blank-lines): New function that tallies up the number of
blank and whitespace-only lines in the current input. One downside of
this design is that this function's conclusions aren't shared with
`erc--discard-trailing-multiline-nulls', which must decide on its own
how many lines to strip.
(erc--check-prompt-explanation): New variable.
(erc--check-prompt-input-for-multiline-blanks): Rework significantly
to provide more informative messages and more sensible behavior for
common cases with respect to relevant option values.
(erc--check-prompt-input-functions): Add new validation function
`erc--check-prompt-for-something'.
(erc--run-input-validation-checks): Set `abortp' slot of
`erc--input-split' when hooks return a non-string, rather than
generate an unhelpful fallback message. Also print a message when
the variable `erc--check-prompt-explanation' is non-nil.
(erc-send-current-line): When the `abortp' slot of the current
`erc--input-split' object is non-nil, forgo normal input processing.
This fixes a regression in 5.6-git, which previously emitted an error
message when it should have just exited the command.
(erc--discard-trailing-multiline-nulls): Always run, regardless of
`erc-send-whitespace-lines', and leave a blank line behind when
stripping a message consisting of only blank lines.
(erc--run-send-hooks): Always run hooks and surrounding logic rather
than only when hooks are populated.
* test/lisp/erc/erc-tests.el (erc--refresh-prompt): Add assertion and
use default value for `erc--input-review-functions'.
(erc-ring-previous-command): Use default value for
`erc--input-review-functions' and ensure `erc-current-nick' does not
error.
(erc--blank-in-multiline-input-p): Remove obsolete test.
(erc--check-prompt-input-functions): Update expected message.
(erc--discard-trailing-multiline-nulls, erc--count-blank-lines): New
tests.
(erc-tests--check-prompt-input--expect,
erc-tests--check-prompt-input-messages): New helper variables.
(erc--check-prompt-input-for-multiline-blanks,
erc--check-prompt-input-for-multiline-blanks/explanations): New tests.
(erc-send-whitespace-lines): Expect hook-processing logic to pad empty
lines instead of deferring to `erc-send-input-line-function'.
(Bug#66073)
F. Jason Park [Tue, 19 Sep 2023 05:50:28 +0000 (22:50 -0700)]
Prefer ticks/hz pairs for some ERC timestamps on 29+
* lisp/erc/erc-compat.el (erc-compat--current-lisp-time): New macro to
prefer ticks/hz pairs on newer Emacs versions because stamps of this
form are easier to compare at a glance when used as values for text
properties.
* lisp/erc/erc-stamp.el (erc-stamp--current-time): Use compat macro.
(Bug#60936)
F. Jason Park [Thu, 21 Sep 2023 04:40:36 +0000 (21:40 -0700)]
Exempt text-scale-mode from erc-scrolltobottom-all
* lisp/erc/erc-goodies.el (erc--scrolltobottom-post-ignore-commands):
New variable, a list of commands that should not trigger a re-scroll.
(erc--scrolltobottom-on-post-command): Don't `recenter' when the
current command appears in `erc--scrolltobottom-post-ignore-commands'.
This fixes a bug that prevented managed recentering when disabling
`text-scale-mode'. Thanks to Corwin Brust for reporting this. Note
that this doesn't address any fill-related alignment issues involving
`text-scale-mode'. Nor does it address likely problems involving the
erroneous restoration of point after the insertion of invisible lines.
(erc-keep-place-enable, erc-keep-place-mode): Adjust hook depth to
fall within reserved interval.
(erc-keep-place-indicator-enable, erc-keep-place-indicator-mode,
erc--keep-place-indicator-on-global-module): Lower hook depth to fall
within reserved interval. Note that this hook has a buffer-local
value, so all members run after their global counterparts.
(Bug#66073)
F. Jason Park [Tue, 19 Sep 2023 05:50:28 +0000 (22:50 -0700)]
; Increase ERC test server queue size
* test/lisp/erc/erc-scenarios-scrolltobottom.el
(erc-scenarios-scrolltobottom--normal,
erc-scenarios-scrolltobottom--all): Use updated name for test fixture.
* test/lisp/erc/resources/erc-d/erc-d.el
(erc-d--initialize-client): For lengthy batches, `erc-d--filter' may
run multiple times before `erc-d--on-request' can pull from the queue,
which results in discarded incoming messages and match failures.
(erc-d--m, erc-d--log): Convert to ordinary functions.
* test/lisp/erc/resources/erc-scenarios-common.el
(erc-scenarios-scrolltobottom--normal,
erc-scenarios-common-scrolltobottom--normal): Rename test fixture from
former to latter and attempt to fix intermittent failure re
`count-screen-lines'.
F. Jason Park [Mon, 25 Sep 2023 13:06:13 +0000 (06:06 -0700)]
Fix wrong "User" column count in erc-ibuffer
(ibuffer-make-column-erc-members): Show tally of all server users for
non-target buffers, and show correct count for targets. Thanks to
Mattias Engdegård for reporting this.
(erc-ibuffer-limit-map): Use "new" `define-ibuffer-filter'
API. (Bug#66191)
F. Jason Park [Mon, 2 Oct 2023 23:29:28 +0000 (16:29 -0700)]
Remove newline from erc-fill regexp
* lisp/erc/erc-fill.el (erc-fill): Remove newline from character
alternative in pattern for skipping past blank and whitespace-only
lines. It seems that as of e61a0398 "regex.c: Consolidate the two
analysis functions", Emacs no longer sees a newline-dollar sequence as
matching an empty line. Also lose `save-match-data', which doesn't
appear to serve any purpose here.
Po Lu [Mon, 2 Oct 2023 07:38:30 +0000 (15:38 +0800)]
Properly translate touch screen events into mouse drag ones
* doc/lispref/commands.texi (Drag Events): Correct misleading
example form and reword subsequent elaboration.
* lisp/touch-screen.el (touch-screen-handle-point-up)
<mouse-drag>: Set posns representing drag start and terminus to
mouse position lists, in lieu of pairs between windows and posns
employed previously.
* test/lisp/eshell/esh-proc-tests.el
(esh-proc-test/pipeline-connection-type/middle)
(esh-proc-test/pipeline-connection-type/last): Use '(ignore)', since
that causes no output when called with no arguments, thus avoiding a
risky 'process-send-string'.
Stefan Kangas [Sun, 1 Oct 2023 17:00:26 +0000 (19:00 +0200)]
; Add several "duplicate value in `cond`" FIXMEs
* lisp/gnus/gnus-cloud.el (gnus-cloud-decode-data):
* lisp/org/org.el (org-read-date-analyze):
* lisp/org/ox-html.el (org-html-latex-fragment):
Add several "duplicate value in `cond`" FIXMEs.
Found by Mattias Engdegård <mattias.engdegard@gmail.com>.
Ref: https://debbugs.gnu.org/51368#51
Stefan Kangas [Sun, 1 Oct 2023 13:37:06 +0000 (15:37 +0200)]
Improve documentation of `ns-use-proxy-icon`
* doc/emacs/macos.texi (Mac / GNUstep Customization):
* src/nsfns.m (syms_of_nsfns): Fix documentation of
'ns-use-proxy-icon'; explain what it does and how it's
used. (Bug#66190)
projects / emacs.git / log