It's important to adjust the match data in between calling
before-change-functions and after-change-functions, so that buffer
change hooks will always see match-data consistent with buffer content.
(Bug #23917)
* src/insdel.c (replace_range): Add new parameter ADJUST_MATCH_DATA, if
true call update_search_regs. Update all callers (except
Freplace_match) to pass 0 for the new parameter.
* src/search.c (update_search_regs): New function, extracted from
Freplace_match.
(Freplace_match): Remove match data adjustment code, pass 1 for
ADJUST_MATCH_DATA to replace_range instead.
Backport from master (Bug#24033).
Inspired by a suggestion by Florian Weimer in:
https://sourceware.org/ml/libc-alpha/2016-07/msg00425.html
* src/emacs.c (main) [__PPC64__]:
Special case for __PPC64__, which needs ASLR disabled in
dumped Emacs too.
This expansion of the FIXME is so that future developers are aware of
the potential problems of aliasing fboundp to symbol-function without
taking backwards compatibility into account.
* src/data.c (fboundp): Note potential backwards compatibility issues in
FIXME.
Copyright-paperwork-exempt: yes
* src/editfns.c (Ftranspose_regions): Regions may overlap, so
use memmove instead of memcpy (bug#19213).
Copyright-paperwork-exempt: yes
This is a backport from trunk.
(cherry picked from commit 354f9f0fc6)
Make the documentation of the search functions more accurate,
complete, and uniform; in particular, extend the description of
the effect when the 'count' parameter is a negative number to all
of these functions.
* src/search.c (Fsearch_backward, Fsearch_forward)
(Fre_search_backward, Fre_search_forward)
(Fposix_search_backward, Fposix_search_forward):
* lisp/isearch.el (word-search-backward, word-search-forward)
(word-search-backward-lax, word-search-forward-lax): Improve doc
strings as described above.
* doc/lispref/searching.texi (String Search, Regexp Search)
(POSIX Regexps): Use 'count' instead of 'repeat' as the name of
the fourth parameter of the *-search-{forward,backward} functions
and improve documentation as described above.
* lisp/simple.el (undo-auto--undoable-change-no-timer): New function.
* src/fileio.c (insert-buffer-contents): Call
undo-auto--undoably-changed-buffers before changes.
Addresses Bug#23785.
Do not merge to master as c98bc98 also addresses the same bug.
Problem reported by Christer Ekholm (Bug#23864).
Backport from master.
* src/process.c (wait_reading_process_output):
Fix typo introduced in 2015-07-06T02:19:13Z!eggert@cs.ucla.edu
when wait == INFINITY and got_output_end_time is invalid.
In this case the code should break, not continue.
* src/undo.c (record_insert): Use record_point instead of
prepare_record, and do so unconditionally.
(prepare_record): Do not record first change.
(record_point): Now conditional on state before the last command.
(record_delete): Call record_point unconditionally.
(record_property_change): Use prepare_record.
(record_marker_adjustments): Use prepare_record.
Addresses Bug# 21722
* src/search.c (Freplace_match): Error out if buffer modification
hooks triggered by buffer changes in replace_range, upcase-region,
and upcase-initials-region clobber the match data needed to be
adjusted for the replacement. (Bug#23869)
* src/dispextern.h (FONT_TOO_HIGH): Don't consider a font "too
high" if its pixel_size value is zero. This avoids assertion
violations at the end of x_produce_glyphs.
* src/w32.c (init_environment): Improve warning message that pops when
Emacs sets HOME according to existence of C:\.emacs (Bug #11612).
Co-authored-by: Eli Zaretskii <eliz@gnu.org>
* src/minibuf.c (Fassoc_string): Clarify how CASE-FOLD affects the
string comparison. (Bug#23833)
* src/fns.c (Fcompare_strings): Fix the description of how
IGNORE-CASE affects the comparison.
* doc/lispref/strings.texi (Text Comparison): Clarify how
CASE-FOLD affects the string comparison in 'assoc-string'. Fix
the description of how IGNORE-CASE affects the comparison in
'compare-strings'.
The commit 1fd3172d "(Fbackward_prefix_chars): Set point properly while
scanning" (1998-03-18), moved the check against of the position against the
buffer beginning out the loop condition so that we might end up checking
the syntax of characters before the beginning of the buffer. This can
cause segfaults or trigger a "Point before start of properties" error in
`update_interval' (called indirectly from `char_quoted').
* src/syntax.c (Fbackward_prefix_chars): Stop the loop when beginning of
buffer is reached (Bug #3552, Bug #17132, Bug #19379).
* src/dbusbind.c (Fdbus__init_bus): Use make_save_pointer to store
connection address. (Bug#23741)
(xd_lisp_dbus_to_dbus): Use XSAVE_POINTER to retrieve connection
address.
Problem reported by Vincent Belaïche (Bug#23704).
* src/syntax.c (skip_chars): Recompute pointers into the
buffer after every call to update_syntax_table_forward,
as it can GC.
* src/xfaces.c (Finternal_set_lisp_face_attribute): Check the
FRAME argument before using it. This avoids gratuitous aborts in
Emacs built with --enable-checking when the luser was unlucky
enough to get the argument list in wrong order.
* src/xftfont.c (syms_of_xftfont) <xft-font-ascent-descent-override>:
New variable.
(xftfont_open): Use it to work around problems with rendering some
fonts. (Bug#23360)
Problem reported by Matthias Clasen (Bug#23144).
This was fixed in a different way in master.
Do not merge to master.
* src/xterm.c (handle_one_xevent) [GTK_CHECK_VERSION (3, 20, 3)]:
Do not call xg_frame_resized in the MapNotify case.
* doc/lispref/processes.texi (Process Buffers): Document
'set-process-window-size' and
'window-adjust-process-window-size-function'.
* etc/NEWS: Mention the new functionality and variable.
* src/process.c (Fset_process_window_size): Improve the doc string.
* doc/emacs/text.texi, doc/lispintro/emacs-lisp-intro.texi:
* doc/lispref/control.texi, doc/lispref/display.texi:
* doc/lispref/help.texi, doc/lispref/strings.texi, lisp/subr.el:
* src/callint.c, src/doprnt.c, src/editfns.c:
Document quoting a bit more systematically.
Problem reported by Alan Mackenzie (Bug#23425).
* src/fns.c (Frequire): Mention `load-path' and fill the doc
string (bug#18829).
* src/lread.c (syms_of_lread): Mention that `require' uses
`load-path'.
(cherry picked from commit 3eca9a0381)
* src/minibuf.c (syms_of_minibuf): Mention
minibuffer-depth-indicator-mode in the doc string to
enable-recursive-minibuffers (bug#14147).
(cherry picked from commit 23ba488a5e)
* src/fns.c (sort_vector_copy): Use a different way to attempt to
work around GCC 3.0-and-earlier incompatibility with C99, one that
does not have problems with modern non-GCC compilers.
* config.bat:
* msdos/sedlisp.inp:
* msdos/sedlibmk.inp:
* msdos/sedleim.inp:
* msdos/sedadmin.inp:
* msdos/sed6.inp:
* msdos/sed3v2.inp:
* msdos/sed2v2.inp:
* msdos/sed1v2.inp: Adapt to Emacs 25.
* src/process.c (remove_slash_colon): Move out of "#ifdef
subprocesses" block, as it its called unconditionally. Move
ADD_SUBFEATURE calls into "#ifdef subprocesses" block, as they
reference variables only defined in that block.
* src/msdos.h: Provide prototypes for IT_set_frame_parameters,
faccessat, msdos_fatal_signal, syms_of_msdos, pthread_sigmask,
dos_keysns, dos_keyread, run_msdos_command, and
syms_of_win16select, to avoid compiler warnings.
* src/msdos.c (SYS_ENVIRON): Define to either '_environ' or
'environ', depending on the DJGPP version.
Remove declarations of externally-visible Lisp objects, like
Qbackground_color and Qreverse.
(run_msdos_command): First argument is not signed, not unsigned.
Use SYS_ENVIRON.
(sys_select): Use 'timespec_cmp' instead of 'timespec_sign', as
the latter doesn't work when 'time_t' is an unsigned data type.
This caused idle timers to behave incorrectly: they only fired
after a keyboard input event.
* src/frame.c (adjust_frame_size) [MSDOS]: Account for
FRAME_TOP_MARGIN that isn't counted in the frame's number of
lines, but dos_set_window_size needs it to be added.
* src/lread.c (INFINITY, NAN) [DJGPP < 2.05]: Provide definitions.
* src/fns.c (sort_vector_copy) [__GNUC__ < 4]: Provide a prototype
that works around compilation errors with older GCC versions.
* src/w16select.c: Don't declare QCLIPBOARD and QPRIMARY as Lisp
Objects.
* src/filelock.c [MSDOS]: Ifdef away most of the code. Provide
no-op implementations for 'lock_file' and 'unlock_file'.
(Ffile_locked_p) [MSDOS]: Always return nil. This avoids multiple
ifdefs in all users of filelock.c functionality.
* src/conf_post.h (EOVERFLOW, SIZE_MAX) [DJGPP < 2.04]: Define.
* src/emacs.c [MSDOS]: Include dosfns.h, to avoid compiler
warnings.
* src/dosfns.h: Provide prototypes for dos_cleanup,
syms_of_dosfns, and init_dosfns.
* src/deps.mk (atimer.o): Depend on msdos.h.
(emacs.o): Depend on dosfns.h.
* src/atimer.c [MSDOS]: Include msdos.h, to avoid compiler
warnings.
* lisp/window.el (window--adjust-process-windows): Skip the body
if 'process-list' is not available. This avoids failure to start
up on MS-DOS.
* lisp/vc/diff.el (diff-no-select): Test 'make-process', not
'start-process', as the latter is now available on all platforms.
* lisp/textmodes/ispell.el (ispell-async-processp): Replace
'start-process' with 'make-process' in a comment.
* lisp/term/internal.el (IT-unicode-translations): Modify and add
a few translations to display Info files with Unicode markup. Fix
an ancient off-by-one mismatch error with Unicode codepoints.
* lisp/progmodes/compile.el (compilation-start): Test
'make-process', not 'start-process', as the latter is now
available on all platforms.
* lisp/man.el (Man-build-man-command, Man-getpage-in-background):
Test 'make-process', not 'start-process', as the latter is now
available on all platforms.
* lisp/international/mule-cmds.el (set-coding-system-map): Test
'make-process', not 'start-process', as the latter is now
available on all platforms.
* lisp/eshell/esh-cmd.el (eshell-do-pipelines-synchronously): Doc
fix.
(eshell-execute-pipeline): Test 'make-process', not
'start-process', as the latter is now available on all platforms.
