* src/emacs.c (load_pdump): Don't overwrite the leading
directories of the Emacs executable just because the pdumper file
was not found in the expected directory relative to the binary.
This is needed to support builds with a separate --bindir
configure-time option and native-compilation. (Bug#55741)
* src/dispnew.c (adjust_frame_glyphs_for_frame_redisplay): Make
sure we have valid frame glyph matrices for the interactive
session. (Bug#55760)
(adjust_frame_glyphs): Add assertions for when we fail to allocate
valid frame glyph matrices for a TTY frame.
This fixes bug#55684. There, with a minibuffer-only frame at start up,
Emacs tried to switch to this frame, whose selected window was the
mini-window. There is no other active window in this frame, so the
attempt to swith to another window failed.
* src/frame.c (do_switch_frame): On switching to a frame whose selected
window is as above, before selecting the most recently used window, check
this ostensible window is an actual live window. Otherwise leave the
mini-window selected.
This fixes bug #55412. The call to Fselect_window was removed on 2021-03-21
in the commit "Prevent open minibuffers getting lost when their frame gets
deleted". This call is actually needed to determine current elements of the
pertinent window and frame when these are used in the frame title.
* src/frame.c (do_switch_frame): When the selected window in the target frame
is the mini-window, switch away from this window unless there is a valid
minibuffer there.
* src/frame.h (struct frame): Add an incidental comment about the element
selected_window.
* src/minibuf.c (move_minibuffers_onto_frame): No longer set the selected
window of the source frame. This action is now performed later, on returning
to that frame, in do_switch_frame when needed (see above).
* src/xdisp.c (gui_consider_frame_title): Replace the Fselect_window call and
associated ancillary settings.
* lisp/faces.el (face-attribute):
* src/xfaces.c (Fx_family_fonts):
* src/font.c (Ffont_get, Ffont_put): Improve and clarify the doc
strings.
* doc/lispref/display.texi (Low-Level Font): Document the :type
attribute of a font. Improve documentation of 'font-get' and
'font-put'.
(Attribute Functions): Add cross-reference to the description of
face attributes.
* doc/lispref/display.texi (Fontsets):
* src/fontset.c (Fset_fontset_font): Improve and clarify the
documentation of 'set-fontset-font'. Rename the arguments to be
more self-explanatory. (Bug#55086)
* src/xdisp.c (get_overlay_arrow_glyph_row): Don't assume every
character in 'overlay-arrow-string' is one byte long. Reported by
Yuri D'Elia <wavexx@thregr.org>.
This reverts commit 3cccf0a910.
This is a change with far-reaching effects on MS-Windows at the least,
where file-related APIs are shadowed to support transparent support
for UTF-8 encoded file names. Making such changes on a stable branch
for the benefit of a proprietary platform with a 13-year old OS is a
tail wagging the dog. Please don't do that without discussing first.
Use openat only on platforms with O_PATH.
This ports to OS X 10.9 and earlier.
Problem reported by Keith David Bershatsky in:
https://lists.gnu.org/r/emacs-devel/2022-04/msg00805.html
* lib-src/emacsclient.c (local_sockname): Use open, not openat.
* src/sysdep.c (sys_openat): New static function,
which uses openat only if O_PATH is defined.
(emacs_openat): Use it instead of openat.
(emacs_openat_noquit): Remove.
(emacs_open_noquit): Reimplement as per the old emacs_openat_noquit,
but use plain 'open'.
Fix a bug where if GC occurred at the wrong moment when locking a
file, the lock file’s name was trashed so file locking did not work.
This bug was introduced in Emacs 28.1. The bug sometimes caused
filelock-tests-detect-external-change test failures on Fedora 35
x86-64 in an en_US.utf8 locale.
* src/filelock.c (lock_file_1, current_lock_owner, lock_if_free)
(lock_file, unlock_file, Ffile_locked_p):
Use Lisp_Object, not char *, for string, so that GC doesn’t trash
string contents.
(make_lock_file_name): Return the encoded name, not the original.
All callers changed.
This adds back macOS-specific code replaced earlier (bug#48548),
specifically to fix build errors on macOS 10.7.5. See discussion at
https://lists.gnu.org/archive/html/emacs-devel/2022-04/msg00779.html .
* src/sysdep.c (HAVE_RUSAGE_INFO_CURRENT, HAVE_PROC_PIDINFO): New.
(system_process_attributes): Use alternative code or exclude features
when building on older macOS versions.
* src/indent.c (Fvertical_motion): Don't consider fringe bitmaps
as "images" for the purpose of vertical-motion logic dealing with
overshooting buffer positions. (Bug#54946)
* src/xdisp.c (Fwindow_text_pixel_size): Handle the case where
there's a display property at START, and move_it_to overshoots.
Do not merge to master. (Bug#54862)
* src/msdos.h (tcdrain): Redirect to '_dos_commit'.
(openat, fchmodat, futimens, utimensat): Add prototypes.
* msdos/sed1v2.inp (MAKE_PDUMPER_FINGERPRINT): Fix indentation, so
that Make won't consider this line a command.
($(etc)/DOC): Chdir back to ../src, since "make-docfile -d" leaves
us in a wrong directory.
* msdos/sedlibmk.inp (GL_GNULIB_GETRANDOM, GL_GNULIB_MEMMEM)
(GL_GNULIB_SIGDESCR_NP): Define to 1, to get the prototypes from
Gnulib headers.
posix_spawn was restoring the wrong signal set, which still had
SIGCHLD and SIGINT masked, causing problems with child processes that
spawned child processes. (Bug#54667)
See the thread ending at
https://lists.gnu.org/archive/html/emacs-devel/2022-03/msg00067.html
for more details.
* src/callproc.c (emacs_spawn): Pass oldset parameter.
(emacs_posix_spawn_init_attributes): Use correct oldset.
(emacs_posix_spawn_init): Remove intermediate function.
(cherry picked from commit 8103b060d8)
* doc/lispref/commands.texi (Command Modes): Note interaction with
native-compile (bug#54437).
* src/data.c: Add comment about not being supported.
Do not merge to master.
* src/xdisp.c (prepare_menu_bars): Include in the windows passed
to pre-redisplay-functions windows whose point was moved from the
last recorded position. (Bug#54450)
* src/dispextern.h (gui_define_fringe_bitmap): Add prototype.
(max_used_fringe_bitmap): Add declaration.
* src/fringe.c (gui_define_fringe_bitmap): New function.
* src/w32term.c (w32_draw_fringe_bitmap):
* src/xterm.c (x_draw_fringe_bitmap) [USE_CAIRO]: Call
'gui_define_fringe_bitmap' if the terminal-specific bitmap data is
not available when a fringe bitmap is about to be drawn. Don't
try to draw a bitmap that is not known to fringe.c. (Bug#54183)
* src/bidi.c (bidi_find_bracket_pairs): Disable BPA optimization
when there are no strong directional characters inside the
bracketed pair. (Bug#54219)
* src/bidi.c (bidi_resolve_brackets): Fix implementation of UBA's
N0 rule when there are no strong directional characters inside the
bracketed pair. (Bug#54219)
* src/xdisp.c (redisplay_internal): Disable redisplay
optimizations that consider just the current line, when
'display-line-numbers-mode' is turned on in the buffer.
(Bug#54091)
* src/xterm.c (handle_one_xevent): Prevent a signal inside
`setup_coding_system' which crashes recent versions of GLib if
the locale coding system is invalid.
Do not merge to master.
* doc/lispref/display.texi (Attribute Functions)
(Face Attributes): Clarify that the :font attribute of a face and
the font returned by 'face-font' are by default for ASCII
characters. (Bug#53664)
* src/eval.c (signal_or_quit): Bind the correct variable, Qdebugger (not
Vdebugger) to Qdebug in the section for errors in batch jobs.
(syms_of_eval): New DEFSYM for Qdebugger.
* src/filelock.c (lock_file): Move call of file name handler to
`Flock_file'. Determine lock_filename only in case
create_lockfiles is non-nil. Adapt the rest of the function accordingly.
(Flock_file): Do not check for create_lockfiles. Call file name
handler if appropriate. (Bug#53207)
* src/xterm.c (handle_one_xevent): Handle setting of variable
'x_set_frame_visibility_more_laxly' when receiving an Expose or
FocusIn event (Bug#53298).
(Qexpose): Define symbol.
(x_set_frame_visibility_more_laxly): New Lisp variable.
* etc/PROBLEMS: Mention frame redraw problem with the
Enlightenment WM and 'x-set-frame-visibility-more-laxly'
workaround.
