* src/editfns.c (save_excursion_save): New arg PDL,
specifying where to save the state. All uses changed.
(save_excursion_restore): Args are now the marker and info
rather than a pointer to a Lisp_Save_Value containing them.
All uses changed.
* src/eval.c (default_toplevel_binding, Fbacktrace__locals):
Treat excursions like other miscellaneous pdl types.
(record_unwind_protect_excursion): Save data directly
into the pdl rather than creating an object on the heap.
This avoids the need to allocate and free an object.
(do_one_unbind, backtrace_eval_unrewind):
Unwind excursions directly.
(mark_specpdl): Mark excursions directly.
* src/lisp.h (SPECPDL_UNWIND_EXCURSION): New constant.
(union specbinding): New member unwind_excursion.
* src/macfont.m (macfont_descriptor_entity): Use cons instead
of make_save_ptr_int, as this avoids the need for a special
type and function for this one-off.
* src/ftfont.c (struct ftfont_cache_data): New member face_refcount.
(ftfont_lookup_cache): Clear it when initializing.
Use make_mint_ptr, since this typically avoids the need to
allocate a Lisp_Save_Value as refcount is now stored elsewhere.
(ftfont_open2, ftfont_close): Manipulate the reference
count in the struct, not in the save object.
* src/term.c (struct tty_pop_down_menu): New type.
(tty_pop_down_menu, tty_menu_show): Use it, along with
record_unwind_protect_ptr, to avoid allocating a Lisp_Misc.
* src/xmenu.c (struct pop_down_menu): New type.
(pop_down_menu, x_menu_show): Use it, likewise.
* src/xterm.c (x_cr_destroy, x_cr_export_frames):
Use record_unwind_protect_pointer to avoid possibly allocating
a Lisp_Misc.
* src/fileio.c (union read_non_regular): New type.
(read_non_regular, Finsert_file_contents):
Use it to avoid allocating a Lisp_Misc.
* src/keymap.c (union map_keymap): New type.
(map_keymap_char_table_item, map_keymap_internal):
Use it to avoid allocating a Lisp_Misc.
* src/emacs-module.c (init_module_assertions): Just use NULL
instead of allocating a dummy on the stack and then using
eassert. Practical platforms check for null pointer
dereferencing nowadays, so this is good enough.
* doc/lispref/text.texi (Parsing JSON): Describe new :null-object
and :false-object kwargs to json-parse-string and
json-parse-buffer.
* src/json.c
(struct json_configuration): New type.
(json_to_lisp): Accept a struct json_configuration* param.
(json_parse_args): Rename from json_parse_object_type.
(Fjson_parse_string): Rework docstring.
(Fjson_parse_string, Fjson_parse_buffer): Update call to
json_to_lisp.
(syms_of_json): Two new syms, QCnull_object and QCfalse_object.
* test/src/json-tests.el
(json-parse-with-custom-null-and-false-objects): New test.
* lisp/json.el (json-pretty-print): Force distinction between empty
objects and null.
(json-encode-list): Remove responsibility to print "null" as this
value is not a list.
(json-encode): Give higher precedence to lists so that an empty list
is printed as an empty object, not as "null".
* test/lisp/json-tests.el (test-json-encode): Add many tests to check
the behavior of pretty-printing.
Do not assume that 8-byte alignment suffices for all C objects,
as some platforms require 16-byte alignment for some objects,
and this will start to bite us as time goes on (e.g., if an
Emacs module ever uses an object containing a long
double, which requires 16-byte alignment on x86-64).
Conversely, on !USE_LSB_TAG platforms, do not insist on
aligning Lisp objects to a multiple of 8, as this is not
needed for high-order tag bits.
* src/alloc.c (LISP_ALIGNMENT, MALLOC_IS_LISP_ALIGNED):
New constants.
(XMALLOC_BASE_ALIGNMENT, XMALLOC_HEADER_ALIGNMENT):
Removed. All uses replaced by LISP_ALIGNMENT.
(aligned_alloc, laligned, lmalloc, lrealloc, union aligned_Lisp_Misc)
(maybe_lisp_pointer, pure_alloc):
Use LISP_ALIGNMENT rather than GCALIGNMENT.
(aligned_alloc): Do not worry about an alignment of
LISP_ALIGNMENT when MALLOC_IS_LISP_ALIGNED, as the code never
uses aligned_alloc with alignment == LISP_ALIGNMENT in that case.
(__alignof__): Remove. All uses removed.
(MALLOC_IS_GC_ALIGNED): Remove.
All uses replaced with MALLOC_IS_LISP_ALIGNED.
(vector_alignment): Remove.
All uses replaced with LISP_ALIGNMENT.
* src/alloc.c (mark_maybe_pointer):
* src/emacs-module.c (value_to_lisp_bits):
Do not assume GCALIGNMENT == 1 << GCTYPEBITS, as GCALIGNMENT
is 1 on !USE_LSB_TAG platforms now.
* src/lisp.h (GCALIGNMENT) [!USE_LSB_TAG]: Now 1.
(struct Lisp_Symbol, union vectorlike_header, struct Lisp_Cons)
(struct Lisp_String): Simplify test for verifying alignment.
Avoid user confusion by explicitly stating that enabling a theme does
not imply disabling other themes and that theme load functions are not
theme switch functions.
* lisp/custom.el (load-theme, enable-theme): Clarify docstrings.
Frames can die between the time we generate a focus event and the time
we get around to processing it. Do run after-focus-change-function,
since that's idempotent and we want to make sure not to miss
any changes.
* lisp/frame.el (handle-focus-in, handle-focus-out): Check for dead frames.
* lisp/startup.el (command-line): Re-evaluate delayed-init custom
variables one more time after setting up the initial
window-system. (Bug#30994)
* doc/emacs/custom.texi (Early Init File):
* etc/NEWS: Warn against using early-init.el as a substitute for
.emacs.
* src/eval.c (Fdefvaralias): Call `display-warning' if the alias
target has a non-eq value to the variable being aliased.
* test/src/eval-tests.el (defvaralias-overwrite-warning): New test.
(sql-product-alist): Add MariaDB entry
(sql-mariadb-program, sql-mariadb-options, sql-mariadb-login-params,
sql-mode-mariadb-font-lock): New variables, aliases of the MySQL
equivalents
(sql-mariadb, sql-comint-mariadb): New interaction mode functions
for MariaDB
(sql-mode-mysql-font-lock-keywords): Updated font-lock for MySQL
and MariaDB
* lisp/net/soap-client.el: Bump version to 3.1.4.
(soap-type-of): New function.
(soap-resolve-references, soap-decode-type)
(soap-encode-attributes, soap-encode-value): Replace aref
calls with calls to soap-type-of.
* lisp/net/soap-inspect.el (soap-sample-value, soap-inspect):
Replace aref calls with calls to soap-type-of.
Co-authored-by: Noam Postavsky <npostavs@gmail.com>
Instead of working around the behavior delete-frame-functions, just
add an after-delete-frame-functions hook.
* doc/lispref/frames.texi (Deleting Frames): Document
`after-delete-frame-functions'.
* etc/NEWS: Mention `after-delete-frame-functions'.
* lisp/frame.el (blink-cursor--should-blink):
(blink-cursor--rescan-frames, blink-frame-mode): Get rid of the
ugly ignored-frame parameter and switch from
`delete-frame-functions' to `after-delete-frame-functions'.
* src/frame.c (syms_of_frame): New variable
`after-delete-frame-functions'.
(delete_frame): Use it.
* lisp/frame.el (blink-cursor--should-blink): New function.
(blink-cursor-check): Call it.
(blink-cursor--rescan-frames): New function.
(blink-cursor-mode): Wire up `blink-cursor--rescan-frames`; stop
using `focus-in-hook' and `focus-out-hook'.
focus-in-hook and focus-out-hook don't accurately reflect actual
user-visible focus states. Add a new focus interface and mark the old
one obsolete.
* doc/lispref/frames.texi (Input Focus): Document new focus
functions. Remove references to the now-obsolete focus hooks.
* lisp/frame.el (frame-focus-state): New function.
(after-focus-change-function): New variable.
(focus-in-hook, focus-out-hook): Move to lisp from C;
mark obsolete.
* lisp/term/xterm.el (xterm-translate-focus-in)
(xterm-translate-focus-out): Track tty focus in `tty-focus-state'
terminal parameter; call `after-focus-change-function'.
(xterm--suspend-tty-function): New function.
* src/frame.c (Fhandle_switch_frame): Update docstring; don't call
focus hooks.
(focus-in-hook, focus-out-hook): Remove: moved to lisp.
(syms_of_frame): Remove unread_switch_frame; add
Vunread_switch_frame.
* src/keyboard.c:
(Finternal_handle_focus_in): New function.
(make_lispy_event): Always report focus events to lisp; don't
translate them to switch events sometimes. Lisp can take care of
creating synthetic switch-frame events via
`internal-handle-focus-in'.
* src/w32term.c (x_focus_changed): Remove switch-avoidance logic:
just directly report focus changes to lisp.
* src/xterm.c (x_focus_changed): Remove switch-avoidance logic:
just directly report focus changes to lisp.
3434edc731 Enlarge DUMPED_HEAP_SIZE for 64-bit Windows builds
36bbdfc017 Update Unicode data files to version 11.0.0 of Unicode
b7b7a5f4f3 * etc/NEWS: Belatedly call out vc-hg changes in v26.1. (B...
5b6f8b54d1 Clarify the documentation of 'dired-recursive-deletes'
9db97b49cd ; * etc/DEBUG: Add information about debugging libXft prob...
0214ffbe60 Clarify doc string of 'update-glyphless-char-display'
ef35d405b1 Clarify subtle issues with 'eq' in byte-compiled code
c6ef3c8321 Make cl-print respect print-quoted (bug#31649)
26b52ac40e Fix unexpected jumps of window-point in 'set-window-config...
4af077ab4d * etc/emacs.appdata.xml: Update Emacs screenshot.
e5ab25deae Fix cursor movement by 'next-logical-line' after 'next-line'
d20beef5f1 Fix prompt in bookmark.el (Bug#24726)
c57e7eaae8 Improve documentation of 'empty' whitespace-style
# Conflicts:
# etc/NEWS
The following commits were skipped:
acaebed014 ; * src/ftfont.c (ftfont_spec_pattern): Fix whitespace.
97d61f878e Port FC_COLOR change to older fontconfig
55c9bb9f3c Fix comint-get-old-input-default for output field case (Bu...
26819cd1c0 ; ChangeLog.3: Fix typo.
e35a08ea4b Prevent infloop in 'delete-trailing-whitespace'
* lisp/progmodes/cperl-mode.el:
* lisp/progmodes/cc-engine.el:
* lisp/progmodes/cc-mode.el: Fix tabs mixed with space preventing
commit hook from succeeding.
* src/keyboard.c (read_char): Use xevent_start in a couple
more places where it is safe. This is needed with
--enable-gcc-warnings --enable-checking on Fedora 28 x86-64.
Instead of using __builtin_assume_aligned (P, GCALIGNMENT) to
tell GCC that P has alignment 8, use (T *) P where T is the
type of the pointed-to object, to tell GCC that P has native
alignment. This is simpler, matches the intent better, and
should help simplify future improvements. Some of these
changes are to pacify gcc -Wnull-dereference, since GCC is
smarter about pointers now that Emacs no longer uses
__builtin_assume_aligned; these minor changes should improve
code efficiency slightly. On Fedora 28 x86-64 with default
optimization this patch shrinks the size of the Emacs text
segment by 0.36%.
* src/conf_post.h (__has_builtin, __builtin_assume_aligned):
Remove; no longer used.
* src/dbusbind.c (XD_OBJECT_TO_DBUS_TYPE):
Pacify -Wnull-dereference by using XCAR instead of CAR_SAFE
and XCDR instead of CDR_SAFE when this is safe.
* src/fileio.c (Fexpand_file_name):
* src/font.c (clear_font_cache):
Pacify -Wnull-dereference by removing unnecessary NILP test.
* src/keyboard.c (xevent_start): New function.
(read_char, read_key_sequence): Pacify -Wnull-dereference by
using xevent_start instead of EVENT_START.
* src/lisp.h (lisp_h_XUNTAG): Remove; XUNTAG is always a macro
now, since it can no longer be implemented as a function.
(XUNTAG): New third argument CTYPE. All uses changed.
Cast result to CTYPE * instead of using __builtin_assume_aligned.
Simplify by using LISP_WORD_TAG.
(LISP_WORD_TAG): New macro.
(TAG_PTR): Use it.
* src/menu.c (x_popup_menu_1):
Pacify -Wnull-dereference by using XCAR instead of Fcar and
XCDR instead of Fcdr where this is safe.
* lisp/textmodes/ispell.el (ispell--call-enchant-lsmod): Cope with a
version suffix on the binary name, so enchant-2 is converted to
enchant-lsmod-2, not enchant-2-lsmod. (Bug#31761)
(cherry picked from commit a402d9aacb)
* src/keyboard.c (READ_KEY_ELTS): New constant.
(keyremap_step, read_key_sequence): Omit BUFSIZE arg, since it's
always READ_KEY_ELTS. All callers changed.
(grow_bool_vector): Remove; no longer needed.
(read_key_sequence): Use a bool array instead of a Lisp bool
vector, since it's small and this puts less pressure on the GC.
* src/json.c (lisp_to_json_toplevel_1): Fix pointer misuse not
caught by C type checking (json_t ** converted to void * where
the program expected json_t *). Bug caught on Fedora 28 x86-64 via
'./configure CFLAGS="-g3 -O2 -fsanitize=address" CANNOT_DUMP=yes'.
Avoid similar problems in the future by rewriting to use
json_t * instead of json_t **.
read_key_sequence can, in various circumstances, play back recorded
events. Make sure that we set last_nonmenu_event as if we weren't
replaying. Without this change, we leave last_nonmenu_event set to
whatever it was before we started replaying, leading to spurious
random keymap menu prompts appearing after reading terminal control
sequences, the translation of which sometimes causes event replays.
* src/keyboard.c:
(grow_bool_vector): New function
(read_key_sequence): Remember menu event history per-event.
The existing debug print commands don't work in rr, since they touch
stderr. The new xfmt command just calls Fformat and doesn't touch the
stdio streams.
* src/.gdbinit:
(xfmt): New GDB command.
* src/print.c:
(debug_format): New function.
* src/lread.c (read_integer): Fix off-by-1 buffer overrun
introduced in 2018-04-17T23:23:16Z!eggert@cs.ucla.edu. The
bug could occur when Emacs read radixed integers containing
more than 100 digits. Bug caught by AddressSanitizer.
Cease using the long obsolete c++-template-syntax-table.
* lisp/progmodes/cc-align.el (c-lineup-template-args): Cease using
c++-template-syntax-table.
* lisp/progmodes/cc-engine.el (c-beginning-of-inheritance-list)
(c-search-decl-header-end, c-beginning-of-decl-1, c-end-of-decl-1)
(c-guess-continued-construct, c-guess-basic-syntax): Cease using
c++-template-syntax-table.
(c-guess-basic-syntax): Add CASE 5D.6 to handle C++11's "using" type
definition.
* lisp/progmodes/cc-langs.el (c++-make-template-syntax-table)
(c++-template-syntax-table): Remove.
(c-equals-type-clause-kwds, c-equals-type-clause-key): New language
constants/variables.
We used to treat the start of a focus-in, focus-out, and the start of
a paste sequence as normal events bound in global-map, but this
approach produces problems when we recognize events in the middle of
actions that don't immediately dispatch to the command loop.
Now we handle these events internally inside read-key, translating the
focus events to nothing and paste-start into an xterm-paste event that
neatly encapsulates the entire paste sequence.
* lisp/term/xterm.el:
(xterm-paste): Accept an event argument; insert text from event.
(xterm-translate-focus-in,xterm-translate-focus-out)
(xterm-translate-bracketed-paste): New functions.
(xterm-handle-focus-in,xterm-handle-focus-out): Remove.
(xterm-rxvt-function-map): Bind new translation functions.
* lisp/progmodes/cc-engine.el (c-beginning-of-statement-1): At the end, accept
"." as a unary operator (which it now is in brace lists in, e.g., C Mode).
(c-guess-basic-syntax): Just before CASE 9C, move back to the previous brace
list entry in the block, rather than to the first one.
* src/systhread.c (sys_thread_create): Change return type to bool.
Check for errors returned by pthread_attr_setstacksize and
pthread_attr_destroy.
(sys_mutex_init): Abort on errors. Enable mutex checks when checking
is enabled.
(sys_cond_init): Abort on errors.
(sys_mutex_lock, sys_mutex_unlock, sys_cond_wait)
(sys_cond_signal, sys_cond_broadcast, sys_cond_destroy): Check for
errors in debug mode.
