This patch implements some of my suggestions in Bug#8545,
with further changes suggested by Eli Zaretskii (Bug#43439).
* src/doprnt.c: Improve comments.
(SIZE_BOUND_EXTRA): Now at top level, for parse_format_integer.
(parse_format_integer): New static function, containing some of
the old doprnt. Fix a bug that caused doprnt to infloop on
formats like "%10s" that Emacs does not use. We could simplify
doprnt further if we dropped support for these never-used formats.
(doprnt_nul): New function.
(doprnt): Use it. Change doprnt API to exit when either it finds NUL
or reaches the character specified by FORMAT_END. In the typical case
where FORMAT_END is null, take just one pass over FORMAT, not two.
Assume C99 to make code clearer. Do not use malloc or alloca to
allocate a copy of the format FMTCPY; instead, use a small fixed-size
array FMTSTAR, and use '*' in that array to represent width and
precision, passing them as separate int arguments. Use eassume to
pacify GCC in switch statements.
* src/term.c (mouse_get_xy): Call 'mouse_position' passing it the
value of 'tty-menu-calls-mouse-position-function' as the
argument.
(syms_of_term) <tty-menu-calls-mouse-position-function>: New
DEFVAR_BOOL.
* src/frame.c (mouse_position): New function, with most of the
code from Fmouse_position, but call 'mouse-position-function' only
if called with non-zero argument.
(Fmouse_position): Call 'mouse_position' to do the job.
* lisp/xt-mouse.el (xterm-mouse-translate-1): Respect
'track-mouse'.
(xterm-mouse-mode): Set 'tty-menu-calls-mouse-position-function'
when setting 'mouse-position-function'.
(xterm-mouse-tracking-enable-sequence): Use SET_ANY_EVENT_MOUSE
(0x1003) so that mouse movement can be reported even if no buttons
are pressed. Doc fix.
* lisp/menu-bar.el (menu-bar-define-mouse-key): New function.
(tty-menu-navigation-map): Call it.
* doc/lispref/frames.texi (Mouse Position): Document
'tty-menu-calls-mouse-position-function'.
* etc/NEWS: Announce 'tty-menu-calls-mouse-position-function'.
* src/image.c (svg_load_image): Calculate the image size by using the
viewBox size and applying it to the image.
* etc/PROBLEMS: Describe the problem with librsvg 2.45 and below.
* src/lread.c (read_escape): Produce better diagnostic for
malformed \u Unicode escapes, while avoiding assertion violation
when READCHAR returns -1 because the input is exhausted.
(Bug#44084)
This is a prerequisite to remove the old C functions, and gives a
measured 3 ms slowdown on my machine, from 0.27s to 0.30s per call to
describe-buffer-bindings (average over 50 calls).
* src/keymap.c (Fdescribe_buffer_bindings): Call Lisp function
describe-map-tree instead of C function describe_map_tree.
The previous conversion of describe_vector from C to Lisp for the
keymap and char table case lead to an unacceptable performance hit.
Moving back to the C version, as we do here, makes this function
around 50 times faster.
The Lisp version of `substitute-command-keys' was benchmarked using
the form `(documentation 'dired-mode)', which now takes less than 8 ms
on my machine. This is around 16 times slower than the previous C
version.
Thanks to Stefan Monnier for helpful pointers on benchmarking.
* src/keymap.c (Fhelp__describe_vector): New defun to expose
describe_vector to Lisp for keymaps and char tables.
(syms_of_keymap): New defsubr for Fhelp__describe_vector.
* lisp/help.el (describe-map): Use above defun instead of Lisp
version.
(help--describe-vector): Remove defun; keep it commented out for now.
Third step in converting substitute-command-keys to Lisp.
* lisp/help.el (describe-map): New Lisp version of describe_map.
(help--describe-map-compare, help--describe-translation)
(help--describe-command, help--shadow-lookup): New helper
functions for describe-map.
(help--keymaps-seen, help--previous-description-column): New
variables.
* src/keymap.c
(Fkeymap__get_keyelt): New defun to expose get_keyelt to Lisp.
(Fdescribe_map_tree_old, Fdescribe_map): Remove defuns.
(Fdescribe_vector_internal): New defun to expose describe_vector to
Lisp in a way usable by describe-map.
(syms_of_keymap): New defsubrs for Fkeymap__get_keyelt and
Fdescribe_vector_internal. Remove defsubrs for Fdescribe_map_tree_old
and Fdescribe_map. Remove 'help--keymaps-seen'.
* test/lisp/help-tests.el
(help-tests-substitute-command-keys/shadow): Extend test.
(help-tests-substitute-command-keys/test-mode)
(help-tests-substitute-command-keys/compare-all)
(help-tests-describe-map-tree/no-menu-t)
(help-tests-describe-map-tree/no-menu-nil)
(help-tests-describe-map-tree/mention-shadow-t)
(help-tests-describe-map-tree/mention-shadow-nil)
(help-tests-describe-map-tree/partial-t)
(help-tests-describe-map-tree/partial-nil): New tests.
This is the second step in converting substitute-command-keys to Lisp.
* lisp/help.el (describe-map-tree): New Lisp version of
describe_map_tree.
(substitute-command-keys): Update to use above function.
* src/keymap.c (Fdescribe_map): New defun to expose describe_map to
Lisp.
* src/keymap.c (syms_of_keymap): New variable 'help--keymaps-seen'; a
temporary kludge planned for removal. New defsubr for Fdescribe_map.
This is only the first step towards a full Lisp implementation, and
does not remove the old C code. On the contrary, it is partly based
on using the old C code, which is to be replaced in steps. This also
makes it easy to test that it produces the same output as the old.
* src/doc.c (Fsubstitute_command_keys_old): Rename from
Fsubstitute_command_keys.
(Fget_quoting_style): New defun to expose text_quoting_style to Lisp.
(syms_of_doc): Expose above symbols.
* lisp/help.el (substitute-command-keys): New Lisp version of
substitute-command-keys. (Bug#8951)
* src/keymap.c
(Fdescribe_map_tree): New defun to expose describe_map_tree to Lisp.
(syms_of_keymap): New defsubr for Fdescribe_map_tree.
* src/keyboard.c (help_echo_substitute_command_keys):
* src/doc.c (Fdocumentation, Fdocumentation_property):
* src/print.c (print_error_message):
* src/syntax.c (Finternal_describe_syntax_value): Fix calls to use new
Lisp implementation of substitute-command-keys.
* test/src/doc-tests.el: Remove file.
* test/lisp/help-tests.el: Add tests for substitute-command-keys
copied from above file.
* src/xterm.c (handle_one_xevent) [USE_GTK]: Don't obey
ConfigureNotify events if the frame is not visible. (Bug#44002)
Copyright-paperwork-exempt: yes
65078e0a76 * lisp/info.el (Info-hide-note-references): Doc fix. (Bug...
30305b543d Make lisp/progmodes/js.el dependent on CC Mode in the Make...
c37b2a9b42 Yet another fix for 'set-minibuffer-message'
72dd911981 Fix posn-at-x-y in builds --without-x
ace25f2066 Clarify the seq-reduce documentation
7d598e281d Make tramp-completion-reread-directory-timeout obsolete (B...
2c31ce18ea Fix 'message' when there's active minibuffer on another frame
# Conflicts:
# doc/misc/tramp.texi
# etc/NEWS
* src/keyboard.c (make_lispy_position): Don't exclude the
window_or_frame = frame case from TTY-only builds. Reported by
Jared Finder <jared@finder.org>.
* doc/lispref/commands.texi (Click Events): Document the format of
POSITION in click events on the frame's internal border.
* src/emacs.c (main): Mark the return from strerror as a constant,
since it shouldn't be changed (bug#43982).
* lib-src/movemail.c (pfatal_and_delete): Ditto.
Copyright-paperwork-exempt: yes
* doc/lispref/numbers.texi (Comparison of Numbers):
Copy some useful text from eql help string.
* src/fns.c (Feql): In doc string, say that eql also compares
integers by value.
* doc/lispref/internals.texi (Module Values): Document
make_unibyte_string (bug#34873).
* src/emacs-module.c (module_make_unibyte_string): New function.
(initialize_environment): Export it.
* src/module-env-25.h: Define it.
* test/data/emacs-module/mod-test.c (Fmod_test_return_unibyte):
Test it.
* test/src/emacs-module-tests.el (module/unibyte): Test it.
Fix mistakes made when removing Cocoa code from nsfont.m.
* src/nsfont.m (nsfont_draw): Remove spurious #ifdef.
(ns_uni_to_glyphs): The #if/#endif was removed from this code, but the
code itself not removed. Remove it now.
c00606171f (origin/emacs-27) A better fix for bug#43886
3196fd44c3 Avoid crashes when a theme is loaded with one frame suspended
0407b15500 Removed an incorrectly placed extra word in the semantic docs
040dcbe53e Fix current-line hscrolling when overlays change
c56eeba2ce Extend tests for shell-command-dont-erase-buffer
On a 32bit build, Emacs can otherwise crash with a !FIXNUM_OVERFLOW_P
assertion in `Ffont_info` by simply doing `emacs -Q` and then `C-s`.
* src/font.c: Try and detect uninitialized `max_width` fields.
(font_make_object): Set max_width to a silly value.
(Ffont_info): Check the value is not silly any more.
* doc/lispref/commands.texi (Command Loop Info):
`this-command-keys' does not include the C-u any more, so remove
that from the description and the example
(bug#22107).
* src/keyboard.c (Fthis_single_command_keys): Don't say that
`this-command-keys' returns the C-u prefix (bug#22111).
This makes it possible to run most of the redisplay code (tho not the
actual drawing since there's nowhere to draw) even when there's no
real frame at hand, as is the case in batch mode.
This makes `xdisp-tests--minibuffer-resizing` work even in batch.
(redisplay_internal): Obey it.
(init_xdisp): Set `echo_area_window` even in noninteractive mode.
* src/dispnew.c (update_frame): Skip the initial frame.
* src/frame.c (make_frame): Use 80x25 as the default initial size.
* test/src/xdisp-tests.el (xdisp-tests--minibuffer-resizing):
Use the new var and fix use of `executing-kbd-macro`.
