* lisp/emacs-lisp/comp.el (comp-loop-insn-in-block): New macro.
(comp-call-optim-func, comp-dead-assignments-func)
(comp-remove-type-hints-func): Use `comp-loop-insn-in-block'.
* src/comp.c (init_gccjit_functions): Use LOAD_DLL_FN_OPT macro to
load gcc_jit_version_major, gcc_jit_version_major and
gcc_jit_version_patchlevel.
* src/w32common.h (LOAD_DLL_FN_OPT): Add macro optionally load a
function from a DLL.
* src/comp.c (cast_kind_of_type): Enum that specifies the kind of type
in the cast enum (unsigned, signed, pointer).
(emit_coerce): Throw an ICE when asked to emit a cast with sign
extension.
(define_cast_from_to): Return NULL for casts involving sign extension.
(define_cast_functions): Specify the kind of each type in the cast
union.
This is to dump prettier C files.
This does not affect compilation times in my tests.
* src/comp.c: Define a 15x15 cast matrix. Use it in emit_coerce().
* src/comp.c (emit_static_object): Use a chunck size of 200 bytes
on bugged GCCs and a longer one (1024) in sane ones. Rename
str in buff to disambiguate and prefer xmalloc to a VLA given
the buffer is not that small.
This change drastically reduce compile time. Apparently GCC optimizer
does not scale up well at all for long sequences of assignments into a
single array.
Nicolás Bértolo <nicolasbertolo@gmail.com>
Andrea Corallo <akrl@sdf.org>
* src/comp.c (gcc_jit_context_new_string_literal)
(gcc_jit_block_add_assignment_op): New imports.
(comp_t): New 'size_t_type' 'memcpy' fields.
(emit_static_object): Define static objects using string literals
and memcpy.
(define_memcpy): New function.
(Fcomp__init_ctxt): Define 'size_t_type' and 'memcpy'.
* src/xdisp.c (Fline_pixel_height, Fmove_point_visually): Save and
restore the bidi cache, to avoid inadvertently resetting it by
starting a new iteration through buffer text. This could cause
trouble if these functions are called during a redisplay cycle,
especially while we were processing RTL text.
Make the navigation in the occur buffer closer
to the navigation in the compilation buffer.
Add bindings to navigate the occur matches (Bug#39121).
Honor `next-error-highlight' and `next-error-highlight-no-select'
when navigating the occurrences.
* lisp/replace.el (occur-highlight-regexp, occur-highlight-overlay):
New variables.
(occur-1): Set `occur-highlight-regexp' to the searched regexp.
(occur-goto-locus-delete-o, occur--highlight-occurrence): New defuns.
(occur-mode-display-occurrence, occur-mode-goto-occurrence):
Use `occur--highlight-occurrence'.
(occur-mode-map): Bind n to `next-error-no-select'
and p to `previous-error-no-select'
* etc/NEWS (Changes in Specialized Modes and Packages in Emacs 28.1):
Announce this change.
* test/lisp/replace-tests.el (replace-tests-with-highlighted-occurrence):
Add helper macro.
(occur-highlight-occurrence): Add test.
loadup.el calls `native-comp-available-p', that calls
load_gccjit_if_necessary() in Windows. That function tries to load
libgccjit using the mappings defined in `dynamic-library-alist'. That
mapping is filled by term/w32-win.el, but that file may be loaded too
late.
* src/emacs.c (syms_of_emacs): Add libgccjit to the
`dynamic-library-alist' used when starting to dump so
`native-comp-available-p' always works in Windows.
This happened when subr.eln was not the first native compilation unit
to be loaded. register_native_comp_unit() is called when loading a
native compilation unit and that in turn used to call `gensym', which
was not loaded yet. This led to a SIGSEGV.
* src/comp.c (register_native_comp_unit): Replace the call to `gensym'
with an ad-hoc counter.
This fixes the use case when, for example, 'find-file-hooks'
fails.
* lisp/files.el (find-alternate-file): If buffer 'oname' exists,
kill it before renaming the new one. (Bug#41359)
* src/sysdep.c (emacs_exec_file): Don't compile this function
anymore on WINDOWSNT, since it is not used there. This function
was the only reason for having 'execve' prototype in ms-w32.h.
* nt/inc/ms-w32.h (execve): Remove prototype and the MinGW64 vs
ming.org mess that it causes.
Being quite early in startup initialization is better not to rely on
Ffile_exists_p, this call Ffile_expand and not all the necessary
initialization already happened.
* src/pdumper.c (dump_do_dump_relocation): Use fopen instead of
Ffile_exists_p.
* src/xdisp.c (redisplay_internal): Don't use "optimization 1"
if a glyph row from which to start display begins with a display
property that draws into the fringes. (Bug#41584)
* src/xdisp.c (try_window_id): Don't use this optimization if a
glyph row from which to start display begins with a display
property that draws into the fringes. (Bug#41584)
(calculate-lisp-indent): To distinguish code and data when indenting,
introduce the convention that a space between an open paren and
a symbol indicate that this should be indented as a simple data list.
* lisp/ido.el (ido-make-buffer-list-1):
Create a buffer-local binding to improve performance when a lot of
buffers are open (bug#41029).
Copyright-paperwork-exempt: yes
The existing ITRANS scheme did not support some characters and
language quirks like 'chillu's. The Inscript method had errors.
* lisp/language/ind-util.el (indian-mlm-base-table): Add archaic
chars, Mozhi combos; cleanup.
(indian-mlm-mozhi-table): New scheme Mozhi.
* lisp/leim/quail/indian.el (inscript-mlm-keytable): Correct
errors. Add Inscript chillus & zero-width chars, Mozhi scheme.
* etc/NEWS: Mention the changes.
* lisp/international/mule.el (define-coding-system):
Require :pre-write-conversion and :post-read-conversion functions
to leave the match data untouched.
* src/coding.c (Fdecode_coding_string, Fencode_coding_string):
Document functions as match-data-preserving.
Suggested by Stefan Monnier (see bug#41445).
bd7b681dc4 (origin/emacs-27) Tiny texinfo markup fixes
d0dd0e0612 ; Fix more @var/@code mixups in Elisp manual
313dc0439e ; Fix another format-spec typo in the Elisp manual
9d7fd78421 Make next-error behavior a bit more flexible
0691d25295 * etc/NEWS.25: Belatedly announce upcase-dwim and downcase...
df91c94ca8 Fix access to single-byte characters in buffer text
* lisp/simple.el (next-error-no-navigation-try-current):
Extract from the case #2 in next-error-find-buffer (bug#40919).
(next-error-find-buffer-function): Use it as the default.
* src/lread.c (read_vector): If the vector is to short to be for
bytecodes don’t do bytecode processing for it, as the processing
might run past the end of the vector.
* configure.ac: Clarify wording for --with-wide-int help.
* src/pdumper.c (dump_vectorlike_generic):
Do the eassert even if --with-wide-int was specified unnecessarily.
* lisp/progmodes/project.el (project--read-project-list): Simplify the
previous commit by utilizing the optional OMIT-NULLS argument to
'split-string'.
Reported by Ture Pålsson.
* lisp/international/ucs-normalize.el
(ucs-normalize-hfs-nfd-post-read-conversion)
(ucs-normalize-hfs-nfd-pre-write-conversion):
Use save-match-data to avoid match data clobber in normalisation.
* test/lisp/international/ucs-normalize-tests.el
(ucs-normalize-save-match-data): New test.
