* src/fileio.c (Finsert_file_contents): Don't signal an error
about not being able to open the file when we're fetching things
via handlers -- in that case, modtime isn't necessarily set, and
if there's a handler, it should take care of this (bug#42431).
* src/frame.c (do_switch_frame): Move call to `get-mru-window` outside
of the "critical region" where (selected-window) != (frame-selected-window).
* src/minibuf.c (move_minibuffers_onto_frame): Add arg `frame`.
Use it instead of `selected_frame`. Adjust all callers.
* src/lisp.h (move_minibuffers_onto_frame): Adjust declaration.
* lisp/progmodes/make-mode.el
(makefile-imake-font-lock-keywords): Give Imake priority over
CPP keywords, and don't require comments to start at BOL.
(makefile-imake-mode): Set comment-start to some sensible value.
* src/widget.c (get_wm_shell): Return WMShellWidget.
(update_wm_hints, widget_update_wm_size_hints): Return whether
or not size hints changed.
* src/widget.h: Delete `size_switch'.
* src/widgetprv.h (EmacsFramePart): Likewise.
* src/xterm.c (x_wm_set_size_hint): Set size hints the usual way
if Xt did not set them. (bug#58412)
* src/window.c (window_scroll_for_long_lines): New function.
(window_scroll): Call 'window_scroll_for_long_lines' when lines
are very long and truncated on display. Also, disable
'fontification-functions' during scrolling in that case.
* src/xdisp.c (redisplay_window): When recentering the window's
display, go back to the centering position using a simplified
method, if lines in the buffer are very long and truncated on
display.
* src/xterm.c (x_update_opaque_region, show_back_buffer)
(XTbuffer_flipping_unblocked_hook, x_clear_under_internal_border)
(x_after_update_window_line): Remove and add calls to
block_input as required.
* lisp/erc/erc.el (erc-format-message): Although not mentioned in its
log message, commit 5281946fbf "Make
format-spec accept function substitutions" also fixed a bug involving
the `ignore-missing' parameter of `format-spec'. Until now, ERC has
been relying on the old behavior to gracefully handle malformed server
messages. This commit tries to regain that functionality.
* lisp/view.el (view-remove-frame-by-deleting): Remove variable
that hasn't been used since 2011 -- 357f93d245 reworked the
code and stopped using it (bug#58413).
* lisp/progmodes/python.el (python-shell-send-string-no-output):
Save and restore `comint-last-prompt-overlay' or `comint-last-prompt'.
* test/lisp/progmodes/python-tests.el (python-tests-shell-wait-for-prompt):
New helper function.
(python-tests-with-temp-buffer-with-shell): New helper macro.
(python-shell-completion-1, python-shell-completion-native-1)
(python-shell-completion-native-with-ffap-1)
(python-shell-completion-native-with-eldoc-1): New tests (bug#58389).
* src/sqlite.c (sqlite_prepare_errmsg): New function.
(Fsqlite_execute): Use it to get the same error format for both
execute and select.
(Fsqlite_select): Factored out from here.
Without this patch, Ubuntu 22.04.1 x86-64 "make check" reports a
failure in test/src/emacs-tests.el’s
emacs-tests/bwrap/allows-stdout. One can reproduce the bug
without using the Emacs executable, by running this script:
#!/bin/bash
export LC_ALL=C
exec strace -f -o /tmp/tr bwrap --ro-bind / / --seccomp 20 -- \
cat /dev/null 20< lib-src/seccomp-filter-exec.bpf
This script exits with status 159, because "cat" didn’t get
started (it got a SIGSYS signal early on).
The command "journalctl -g SECCOMP" indicated that rseq (syscall
334) was the problem. This syscall is issued by
/lib64/ld-linux-x86-64.so.2 before ‘main’ is called.
There’s another problem with the clone3 syscall, which is used by
pthread_create starting in glibc 2.34. pthread_create is called
by g_child_watch_source_new, which is called by
init_process_emacs.
* lib-src/seccomp-filter.c (main): Allow rseq, clone3. This
causes the test to pass. Perhaps a fancier, more accurate patch
could be written by someone who has the time.
Discussed in Bug#51143.
* lisp/simple.el (execute-extended-command--shorter): Compute a
complete list of `commandp' symbols once. This significantly speeds
up complicated cases while the slowdown of simple cases is still
accetable.
* src/xfns.c (x_window): Avoid extraneous
XChangeWindowAttributes.
* src/xterm.c (x_display_set_last_user_time): New argument
SET_PROPERTY. Do not change the user time property if it is not
true. All callers changed.
(handle_one_xevent): Avoid setting the user time property on
LeaveNotify and XI_Leave events.
Binding it causes errors when some of the enclosed code ends up
loading a `.el` file, and it is unnecessary since the rest of
the code (in `url-file-asynch-callback`) is already careful to
use `insert-file-contents-literally`.
* src/fns.c (HAVE_FAST_UNALIGNED_ACCESS): New.
(Fstring_lessp): Only use word operations where safe, because string
data from purespace may be unaligned.