GitLab

 lib-src/makefile.w32-in ($(BLD)/emacsclientw.exe): Use $(MWINDOWS)
 instead of a literal -mwindows, which is not supported by MSVC.
 nt/nmake.defs (MWINDOWS): Define as "-subsystem:windows -entry:mainCRTStartup".
 Suggested by Fabrice Popineau <fabrice.popineau@supelec.fr>.
 nt/gmake.defs (MWINDOWS): Define as "-mwindows".

 src/search.c (search_buffer): Remove calls to
 r_alloc_inhibit_buffer_relocation, as it is now called by
 maybe_unify_char, which was the cause of relocation of buffer text
 in bug#11519.

 src/charset.c (maybe_unify_char): Inhibit relocation of buffer text
 for the duration of call to load_charset, to avoid problems with
 callers of maybe_unify_char that access buffer text through C
 pointers.
 src/ralloc.c (r_alloc_inhibit_buffer_relocation): Increment and
 decrement the inhibition flag, instead of just setting or
 resetting it.

Fixes: debbugs:11519

 nt/configure.bat (genmakefiles): Move the redirection away from the
 end of the command, to avoid excess whitespace at the end of Make
 variables created at configure time, and also avoid things like
 "FOO1>>config.settings", where "1" gets interpreted as the file
 descriptor and eaten up.  This fixes breakage introduced by the
 last change, without reintroducing the bug fixed by that change.

* desktop.el (desktop-read): Clear previous and next buffers for all
windows and bury *Messages* buffer (bug#11556).

 src/lisp.h [REL_ALLOC]: Add prototypes for external functions
 defined on ralloc.c.
 src/buffer.c [REL_ALLOC]: Remove prototypes of
 r_alloc_reset_variable, r_alloc, r_re_alloc, and r_alloc_free,
 they are now on lisp.h.
 src/ralloc.c (r_alloc_inhibit_buffer_relocation): New function.
 src/search.c (search_buffer): Use it to inhibit relocation of buffer
 text while re_search_2 is doing its job, because re_search_2 is
 passed C pointers to buffer text.

gnus-msg.el (gnus-msg-mail): Ensure that gnus-newsgroup-name is a string so that Gcc works (bug#11514)

 src/msdos.c (internal_terminal_init) <Vwindow_system_version>:
 Update value to 24.

 src/xdisp.c (move_it_to): Under MOVE_TO_Y, when restoring iterator
 state after an additional call to move_it_in_display_line_to, keep
 the values of it->max_ascent and it->max_descent found for the
 entire line.
 (pos_visible_p): Revert the comparison against bottom_y to what it
 was in 2012-05-13T18:22:35Z!eliz@gnu.org.

 lisp/mail/sendmail.el (mail-yank-region): Recognize
 rmail-yank-current-message in addition to insert-buffer.  Fixes
 mail-mode's "C-c C-r" that otherwise does nothing when invoked in
 a *mail* buffer created through rmail-start-mail with sendmail as
 mail-user-agent.

 nt/configure.bat: Ensure a space between %var% expansion and
 redirection symbol '>', which breaks when %var% ends in a digit,
 such as 1.

parens around the arg list.  Silly backport.

Fixes: debbugs:11499

Regenerate ldefs-boot.el and AUTHORS.

 src/xdisp.c (pos_visible_p): Fix last change.

Suggested by Martin Rudalics.

* doc/lispref/display.texi (Face Functions): Fix define-obsolete-face-alias.

* doc/lispref/functions.texi (Obsolete Functions): Fix doc for
set-advertised-calling-convention.

* doc/lispref/modes.texi (Mode Help): Fix describe-mode.

* doc/lispref/variables.texi (Variable Aliases): Fix make-obsolete-variable.

* lisp/help.el (describe-mode): Doc fix.

Fixes: debbugs:11267

* commands.texi (Recursive Editing): recursive-edit is a
command.
* compile.texi (Docs and Compilation):
byte-compile-dynamic-docstrings is an option.
* debugging.texi (Invoking the Debugger): debug is a command.
* display.texi (Progress): progress-reporter-update and
progress-reporter-force-update have VALUE argument optional.
(Animated Images): Use non-@code{nil} instead of non-nil.
* files.texi (Format Conversion Round-Trip): Use non-@code{nil}
instead of non-nil.
* frames.texi (Creating Frames): make-frame is a command.
(Input Focus): select-frame is a command.
(Pointer Shape): void-text-area-pointer is an option.
* help.texi (Describing Characters): read-kbd-macro is a
command.
(Help Functions): describe-prefix-bindings is a command.
* markers.texi (Creating Markers): Both arguments of copy-marker
are optional.
* minibuf.texi (Reading File Names): Use @kbd instead of @code.
* modes.texi (Mode Line Variables): mode-line-remote and
mode-line-client are not options.
(Imenu): imenu-add-to-menubar is a command.
(SMIE Indentation Helpers): Use non-@code{nil} instead of
non-nil.
* os.texi (Sound Output): play-sound-file is a command.
* package.texi (Package Archives): Use @key{RET} instead of
@kbd{RET}.
* processes.texi (Signals to Processes): Use @key{RET} instead
of @code{RET}.
(Signals to Processes): signal-process is a command.
* text.texi (Clickable Text): Use @key{RET} instead of
@kbd{RET}.
(Base 64): base64-encode-string is not a command while
base64-decode-region is.
* windows.texi (Switching Buffers): pop-to-buffer is a command.

 src/xdisp.c (pos_visible_p): Don't report a position visible when move_it_to
 stopped at the last line of window, which happens to be scanned
 backwards by the bidi iteration.

 src/xdisp.c (handle_stop): Don't call get_overlay_strings_1 if we
 already have overlays loaded.
 (handle_single_display_spec): Before returning without displaying
 fringe bitmap, synchronize the bidi iterator with the main display
 iterator, by calling iterate_out_of_display_property.
 (iterate_out_of_display_property): Detect buffer iteration by
 testing that it->string is a Lisp string.
 (get_next_display_element): When the current object is exhausted,
 and there's something on it->stack, call set_iterator_to_next to
 proceed with what's on the stack, instead of returning zero.
 (set_iterator_to_next): If called at the end of a Lisp string,
 proceed to consider_string_end without incrementing string
 position.  Don't increment display vector index past the end of
 the display vector.

reference commands.

 src/xdisp.c (handle_single_display_spec): Return 1 for left-margin
 and right-margin display specs even if the spec is invalid or we
 are on a TTY, and thus unable to display on the fringes.  That's
 because the text with the property will not be displayed anyway,
 so we need to signal to the caller that this is a "replacing"
 display spec.  This fixes display when the spec is invalid or we
 are on a TTY.

This bug broke the build on AIX.  Problem reported by Gilles Pion.

the following \r\n using a single `process-send-string', since the
Lotus SMTP server refuses to accept any commands if they are sent
with two `process-send-string's.

Fixes: debbugs:11444

Obey pcomplete-arg-quote-list inside double-quoted args as well.

Fixes: debbugs:11348

* doc/lispref/Makefile.in (mostlyclean): Add some more vol1/2 items.