Make some C symbols static if they don’t need to be extern.
Also, remove a couple of functions that were discovered to be
unused as a result of this process, and mark two extern
functions intended to be usable from GDB.
* src/buffer.c (previous_overlay_change):
* src/composite.c (composition_lisp_table):
* src/fileio.c (file_name_directory):
* src/frame.c (check_tty):
* src/insdel.c (adjust_markers_for_insert):
* src/keyboard.c (unread_switch_frame, read_char)
(requeued_events_pending_p):
* src/sysdep.c (renameat_noreplace) [!HAVE_ANDROID]:
* src/textconv.c (get_conversion_field) [!HAVE_ANDROID]:
* src/treesit.c (TREESIT_BOB_LINECOL, TREESIT_TS_POINT_1_0)
(treesit_buf_tracks_linecol_p, make_treesit_parser)
(make_treesit_node):
* src/xdisp.c (x_y_to_hpos_vpos):
* src/xfaces.c (load_color) [!MSDOS]:
* src/xfns.c (x_real_pos_and_offsets):
* src/xterm.c (x_dnd_do_unsupported_drop):
Now static. Move up if necessary.
* src/coding.c (utf8_string_p):
* src/keyboard.c (detect_input_pending_ignore_squeezables):
Remove; unused.
* src/frame.c (gui_set_alpha):
* src/textconv.c (start_batch_edit, end_batch_edit, commit_text)
(finish_composing_text, set_composing_text)
(set_composing_region, textconv_set_point_and_mark)
(delete_surrounding_text, request_point_update)
(textconv_barrier, replace_text, get_extracted_text)
(get_surrounding_text):
* src/xdisp.c (gui_union_rectangles):
* src/xterm.c (xi_frame_selected_for):
Define only if needed.
* src/treesit.c (treesit_debug_print_parser_list)
(treesit_debug_print_linecol): Declare EXTERNALLY_VISIBLE.
* src/xdisp.c (x_y_to_column_row): New function, defined
only if needed. All external callers to x_y_to_hpos_vpos changed.
* src/nsterm.m ([EmacsWindow initWithEmacsFrame:fullscreen:screen:]):
Add a compile check around setToolbarStyle since it is not available
until macOS 11.0 and a run-time check in case it was compiled on
macOS >= 11.0 but run on an earlier version. (Bug#79951)
* msdos/sed2v2.inp (ALIGNOF_EMACS_INT): Define to 4.
* msdos/sedlibmk.inp (OMIT_GNULIB_MODULE_fchmodat)
(OMIT_GNULIB_MODULE_issymlinkat, OMIT_GNULIB_MODULE_issymlink)
(OMIT_GNULIB_MODULE_lchmod): Define to true.
* src/gmalloc.c (__morecore, __after_morecore_hook)
(__malloc_initialize_hook): Don't declare till stddef.h has been
included for `ptrdiff_t'.
This is mostly a cleanup patch to fix growing discrepancies
between src/lisp.h and configure.ac in terms of how they deduce
how EMACS_INT aligns. The patch includes a static check that the
two methods now agree. It also speeds up ‘configure’ a bit.
* configure.ac (ALIGNOF_INT, ALIGNOF_LONG, ALIGNOF_LONG_LONG):
Remove; no longer used.
(ALIGNOF_EMACS_INT): New macro.
(system_malloc): Set to 'no' more consistently with what’s
in src/lisp.h.
* src/lisp.h (ALIGNOF_EMACS_INT): Do not define here, as config.h
defines it now. Check that config.h’s definition equals the
actual alignof (EMACS_INT).
(USE_USB_TAG): Set more consistently with what’s in configure.ac.
(alignas): Don’t second-guess Gnulib.
* src/alloc.c, src/gmalloc.c: Include <malloc.h> before <stdlib.h>.
This mimics include patterns elsewhere, and avoids a problem if
malloc.h declares realloc (i.e., rpl_realloc) early as extern,
whereas Gnulib stdlib.h declares it later as inline which means it
is extern inline, which clashes with gmalloc’s definition of
rpl_realloc when gmalloc.o is used. Problem discovered when
building Emacs on Ubuntu 25.10 with "./configure CC='gcc -m32'
emacs_cv_struct_alignment=no gl_cv_header_working_stdalign_h=no".
* src/frame.c (store_frame_param): Make sure 'minibuffer',
'parent-frame' and 'delete-before' parameters reference frame on
same terminal. Spare GUI frames when checking root frames of
tty frames and their surrogate minibuffer frames (Bug#79947).
The empty vector and string(s) are immutable, contain no references and
always read as the same objects.
* src/print.c (PRINT_CIRCLE_CANDIDATE_P): Turn macro into...
(print_circle_candidate_p): ...a function, and exclude [] and "".
* test/src/print-tests.el (print-circle): Add test case.
* src/nsterm.m ([EmacsWindow initWithEmacsFrame:fullscreen:screen:]):
Always show the toolbar below the window title. This is needed on
Mac OS 11+ where the toolbar style is decided by the system (which
is unpredictable) and the newfangled "compact" toolbar may be chosen
(which is undesirable). (Bug#79900)
Also, add more commentary about the situation.
* src/systhread.h (SYSTHREAD_ALIGN_ROOM): Use alignof (double),
not alignof (int), to align the room. This means we have
only 8 (not 12) bytes of slop on 32-bit HPPA GNU/Linux.
On this platform, you cannot reliably malloc objects containing
values of type pthread_mutex_t or pthread_cond_t, since malloc
guarantees only 8-byte alignment but these two types require
16-byte alignment. See GCC bug 115750
<https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115750> reported by
Sam James and Emacs bug 79936 <https://bugs.gnu.org/79936>
reported by John Paul Adrian Glaubitz.
* configure.ac (ALIGNOF_MAX_ALIGN_T, ALIGNOF_PTHREAD_COND_T)
(ALIGNOF_PTHREAD_MUTEX_T) [HAVE_PTHREAD]: New C macros.
* src/systhread.c (sys_mutex_init, sys_mutex_lock)
(sys_mutex_unlock, sys_cond_init, sys_cond_wait)
(sys_cond_signal, sys_cond_broadcast, sys_cond_destroy):
Use SYTHREAD_ALIGN_PTR to convert sys_mutex_t and sys_cond_t to
pthread_mutex_t and pthread_cond_t, since runtime conversion is
needed on GNU/Linux HPPA.
* src/systhread.h (SYSTHREAD_ALIGN_ROOM, SYSTHREAD_ALIGN_PTR):
New macros, which affect the generated code only on
unusual platforms like GNU/Linux HPPA.
(sys_mutex_t, sys_cond_t) [HAVE_PTHREAD]: Use them.
* src/dbusbind.c (XD_DBUS_VALIDATE_BUS_NAME)
(XD_DBUS_VALIDATE_PATH, XD_DBUS_VALIDATE_INTERFACE)
(XD_DBUS_VALIDATE_MEMBER): Make these expand to single statements
when they are followed by ‘;’. All uses changed.
* lisp/vc/vc-hooks.el (vc-file-setprop): Use cl-pushnew.
(vc--repo-setprop, vc--repo-getprop, vc--repo-clearprops): New
functions.
* lisp/vc/vc.el (vc--incoming-revision): Cache incoming
revisions.
* src/fns.c (Fput): State that VALUE is returned.
* lisp/treesit.el (treesit-auto-install-grammar): Use the first
writable directory from 'treesit-extra-load-path' for values
'always', 'ask', 'ask-dir' (bug#79862).
* src/treesit.c (syms_of_treesit): Extend docstring.
* lisp/treesit.el (treesit-auto-install-grammar): Add the value 'ask-dir'.
(treesit-ensure-installed): When 'treesit-auto-install-grammar' is
'ask' or 'ask-dir', ask for the directory to install the grammar library
using the first writable directory of 'treesit-extra-load-path' as default.
Also add the provided directory to 'treesit-extra-load-path' afterwards.
* src/treesit.c (treesit-extra-load-path): In the docstring
mention the fact that the first directory is special (bug#79862).
Problem reported by John Paul Adrian Glaubitz (bug#79876).
* src/data.c (Fash): Don’t assume stdc_leading_zeros, which
returns an unsigned integer of unspecified width, returns
a value narrower than EMACS_INT. Also, don’t munge the code to
worry about stdc_leading_zeros (0); unlike GCC’s __builtin_clz,
stdc_leading_zeros works fine on 0. And use a temporary to
avoid a cast.
Problem reported by Helmut Eller (bug#79886).
* src/lisp.h (USE_LSB_TAG): Adjust to alignas changes.
(alignas): For GCC and Clang, prefer __attribute__ ((__aligned__ (A)))
to alignas (A) even if the compiler claims to support the latter.
This works around a bug in GCC 14 i386 C23 alignas.
No need to worry about whether __alignas_is_defined is defined.
We dropped support for MSVC long ago, and nowadays only support
MinGW for the MS-Windows builds of Emacs.
* src/w32term.h:
* src/w32fns.c:
* src/w32.c:
* nt/inc/sys/socket.h:
* nt/inc/stdalign.h:
* nt/inc/ms-w32.h: Remove MSVC-specific code and preprocessor
conditions.
* nt/inc/ms-w32.h: Don't redirect Posix functions to MS _FOO
equivalents, as MinGW headers do that nowadays better, not
necessarily via functions, and do not disable that when __STDC__
is defined.
Port to planned glibc 2.43 (scheduled for February 2026),
which will support qualifier-generic standard functions; see:
https://sourceware.org/git/?p=glibc.git;a=commit;h=cd748a63ab1a7ae846175c532a3daab341c62690
For example, strchr (P, C) will return pointer to const if P is
pointer to const. The idea is to catch dumb programming errors when a
program mistakenly uses strchr to convert a pointer to const
to an unrestricted pointer. This feature is required by C23,
and will be enabled by default in GCC 15.
* src/callint.c (Fcall_interactively):
Respect constness of pointer when calling memchr.
* src/gtkutil.c (xg_get_font):
2nd arg is char *, not const char *.
* src/xfaces.c (parse_float_color_comp): Return bool, not double.
New arg DST. All callers changed. This makes it easier for
callers to use char const *.
(parse_color_spec): Respect constness of pointer when calling strchr.
* lisp/cus-start.el: Make 'abbrev-mode' customizable.
* doc/emacs/abbrevs.texi (Abbrev Concepts):
* etc/NEWS:
* src/buffer.c (syms_of_buffer): <abbrev-mode>: Document
enabling Abbrev mode by default in all buffers by customizing
'abbrev-mode' to a non-nil value.
* java/org/gnu/emacs/EmacsNative.java (setEmacsParams): New arg
UIMODE.
(sendConfigurationChanged): New args DETAIL and UI_MODE.
* java/org/gnu/emacs/EmacsNoninteractive.java (main1): Provide
an undefined UI mode.
* java/org/gnu/emacs/EmacsService.java (EmacsService): New field
uiMode.
(onCreate): Initialize this field at start-up and provide the
same to setEmacsParams.
(onConfigurationChanged): If the UI mode has been altered,
generate a configuration changed event to match.
* src/android.c (android_ui_mode): New variable.
(setEmacsParams): New argument UI_MODE. Initialize the same
from this variable.
* src/androidgui.h (enum android_configuration_changed): New
enum.
(struct android_configuration_changed_event): New field
`DETAIL'. Convert fields providing specifics into a union of
display density information and a UI mode integer.
* src/androidterm.c (handle_one_android_event): Handle both
manners of configuration change events.
(android_term_init): Initialize Vtoolkit_theme from UI mode
provided at start-up.
* src/frame.c (syms_of_frame): Always define Vtoolkit_theme.
Define Qtoolkit_theme_set_functions.
* src/gtkutil.c (xg_update_dark_mode_for_all_displays):
* src/w32term.c (w32_read_socket): Generate special toolkit
theme events, rather than executing hooks directly within the
read_socket callback.
* src/keyboard.c (kbd_buffer_get_event)
<TOOLKIT_THEME_CHANGED_EVENT>: Run Qtoolkit_theme_set_functions
and set Vtoolkit_theme from event->ie.arg.
* src/termhooks.h (enum event_kind): New event
TOOLKIT_THEME_CHANGED_EVENT.
* src/window.c (window_set_parent_and_normal_sizes): When
assigning normal sizes in a combination, take the direction from
its parent's 'horizontal' slot and not from that of the first
window in the combination.
Assuming ATTRIBUTE_PURE means that the function must return,
a few more functions that should not be declared with ATTRIBUTE_PURE.
The GCC manual (and even the C23 standard, with [[reproducible]])
is not clear about this, and it’s better to be safe.
* src/bignum.h (mpz_get_d_rounded):
* src/lisp.h (bignum_to_double):
No longer pure, as it does not return if memory is exhausted.
* src/fns.c (Fproper_list_p): No longer pure, as it does not
return if the user quits.
* src/gnutls.c (Fgnutls_errorp): No longer pure, as it does not
return if it runs into an eassert failure in XSYMBOL_WITH_POS via EQ.
* src/lisp.h (bignum_to_intmax, bignum_to_uintmax, bignum_bufsize):
No longer pure, as it does not return if it runs into an
eassert failure in XBIGNUM via xbignum_val.
Problem for mpz_get_d_rounded reported by Helmut Eller in:
https://lists.gnu.org/r/emacs-devel/2025-11/msg00795.html
* lib-src/make-docfile.c (DEFUN_pure): New constant.
(write_globals, scan_c_stream): Support "attributes: pure".
* src/bignum.h (mpz_get_d_rounded):
* src/data.c (Fsymbolp, Fmodule_function_p, Fintegerp, Fnumberp):
* src/lisp.h (bignum_to_double, bignum_to_intmax)
(bignum_to_uintmax, bignum_bufsize):
Now pure, not const, since they depend on current state.
For example, Fsymbolp now inspects symbols_with_pos_enabled,
and the bignum functions inspect bignum contents in memory.
* src/data.c (Feq):
* src/xfaces.c (Fface_attribute_relative_p):
No longer const, since they might abort when debugging.
* src/pdumper.h (pdumper_object_p, pdumper_cold_object_p)
(pdumper_find_object_type, pdumper_object_p_precise):
These are not const functions. But there is no need to declare
them to be pure, either, as they’re inline so the compiler can
figure it out.
* src/keyboard.c (make_lispy_position): When calculating a
position in the left margin skip any scroll bar on the left of
the associated window (Bug#79846).
* src/conf_post.h: Avoid redirecting '*printf' functions if
'OMIT_CONSOLESAFE' is defined to 1.
* nt/cmdproxy.c:
* nt/ddeclient.c: Define 'OMIT_CONSOLESAFE' to 1 to avoid
redirecting '*printf functions' to stdio-consolesafe replacements.
* nt/Makefile.in (LIBS_ADDPM): Add -lgnu to link 'addpm' against
Gnulib. (Bug#79855)
* src/keyboard.c (adjust_point_for_property): Handle 'display'
property on the same text as 'invisible' property the same as the
overlay with a 'display' property. (Bug#79787)
* src/keyboard.c (adjust_point_for_property): Handle the case
where invisible text has an overlay over it with a replacing
'display' property, which leaves some "trace" of the invisible
text on display. (Bug#79787)
MSVCRT implementation of stdio functions which output
characters one by one fails for CJK double-byte encodings.
Gnulib provides replacement functions which work around
those bugs. This change makes Emacs and lib-src programs
use the replacements when Emacs is linked against MSVCRT.
* src/conf_post.h (fwrite, fprintf, printf, vfprintf, vprintf)
[WINDOWSNT]: Redirect to Gnulib replacements when Emacs is linked
against MSVCRT (as opposed to UCRT). Suggested by Bruno Haible
<bruno@clisp.org>.
* src/process.c (wait_reading_process_output): Process status
changes of sub-processes when called with read_kbd zero and some
"keyboard input" is available, but no output from any sub-process.
(Bug#79777)
* src/xdisp.c (reset_message_log_need_newline): New function.
* src/lisp.h: Add prototype of 'reset_message_log_need_newline'.
* src/print.c (print_error_message): Call
'reset_message_log_need_newline' to prevent message3 from
outputting a newline after "user-error:". (Bug#79840)
