* lisp/server.el (server-start): Use 'locale-coding-system' to
encode messages sent back to the client.
* lib-src/emacsclient.c (main): Print '-error' messages via
'message', not directly via 'fprintf'. This shows the error on
MS-Windows when the client is invoked as 'emacsclientw', since
stderr goes to the bit bucket in that case.
Without this patch, Ubuntu 22.04.1 x86-64 "make check" reports a
failure in test/src/emacs-tests.el’s
emacs-tests/bwrap/allows-stdout. One can reproduce the bug
without using the Emacs executable, by running this script:
#!/bin/bash
export LC_ALL=C
exec strace -f -o /tmp/tr bwrap --ro-bind / / --seccomp 20 -- \
cat /dev/null 20< lib-src/seccomp-filter-exec.bpf
This script exits with status 159, because "cat" didn’t get
started (it got a SIGSYS signal early on).
The command "journalctl -g SECCOMP" indicated that rseq (syscall
334) was the problem. This syscall is issued by
/lib64/ld-linux-x86-64.so.2 before ‘main’ is called.
There’s another problem with the clone3 syscall, which is used by
pthread_create starting in glibc 2.34. pthread_create is called
by g_child_watch_source_new, which is called by
init_process_emacs.
* lib-src/seccomp-filter.c (main): Allow rseq, clone3. This
causes the test to pass. Perhaps a fancier, more accurate patch
could be written by someone who has the time.
Prefer static_assert in just one file for now; the idea is to do
it elsewhere eventually. static_assert is standard (starting with
C23) whereas verify is not, and static_assert can be used even in
pre-C23 files due to Gnulib’s support for it.
* lib-src/seccomp-filter.c: Do not include verify.h.
Prefer static_assert to verify.
C23 is adding the C++ keywords bool, true, and false;
prefer them to <stdbool.h> if they are available.
* admin/merge-gnulib (GNULIB_MODULES):
Add stdbool, which emulates C23 on pre-C23 platforms.
(AVOIDED_MODULES): Remove stdbool; Gnulib has renamed
this module to stdbool-c99 and nobody uses it so it does
not need to be avoided.
* m4/c-bool.m4: New file, from Gnulib stdbool module.
* lib-src/seccomp-filter.c, src/conf_post.h, src/dynlib.h:
* src/emacs-module.c, src/nsterm.m, src/systhread.h:
* test/src/emacs-module-resources/mod-test.c:
Use the C23 style and use bool without including <stdbool.h>.
The Gnulib stdbool module causes config.h to include stdbool.h
on pre-C23 platforms.
* src/emacs-module.h.in:
Don’t include <stdbool.h> if C23 or later, or if
it has already been included.
* lib-src/emacsclient.c (main): Use right length modifier when
printing uintmax_t.
* src/alloc.c (check_pure_size): Use right length modifier when
printing ptrdiff_t.
* lib-src/emacsclient.c (DEFAULT_TIMEOUT): Move out of the
!WINDOWSNT condition, to fix the MS-Windows compilation.
(set_socket_timeout) [WINDOWSNT]: Protect against too-large values
of timeout.
* lib-src/emacsclient.c (DEFAULT_TIMEOUT): New constant.
(timeout): New static variable.
(longopts, shortopts, decode_options, print_help_and_exit): Add new
flag --timeout.
(set_socket_timeout, check_socket_timeout): New helper functions.
(main): Display a status message or exit after Emacs has not responded
for a while, depending on above new --timeout flag. (Bug#50849)
* doc/emacs/misc.texi (emacsclient Options):
* doc/man/emacsclient.1: Document the above new option.
* etc/NEWS: Announce it.
This reverts commit 460f35e96d.
Re-remove the code that scans .el files for docstrings, now
that even `lisp/loaddefs.el` is compiled.
* lib-src/make-docfile.c (scan_file): Don't call `scan_lisp_file`.
(scan_lisp_file, skip_white, read_lisp_symbol, search_lisp_doc_at_eol):
Delete functions.
* lib-src/etags.c (readline_internal): Do not copy a pointer to
freed storage, as that has undefined behavior even if the pointer
is not dereferenced.
(relative_filename): Avoid a backward scan by remembering where
the last slash was. This is a bit faster, and pacifies a GCC
false alarm.
This lets ‘./configure; make’ work on Fedora 36 x86-64 from a Git
checkout without generating false-alarm warnings.
* lib-src/etags.c (main): There appeared to be false alarm with
GCC 12. However, the code was wrong anyway, as it mishandled file
names containing "'" so fix that bug. This pacifies GCC.
(mercury_decl): Omit tests ‘s + pos != NULL’ that were apparently
intended to be ‘s[pos] != '\0'’ but which were miscoded to always
be true and which were mostly not needed anyway. In one place,
though, a test was needed, so fix that by using strchr instead.
* src/alloc.c (lisp_free) [!GC_MALLOC_CHECK]:
* src/term.c (Fsuspend_tty): Don’t look at a pointer after freeing
it, even just to test it for equality with some other pointer, as
this has undefined behavior in C and GCC 12 diagnoses this.
* src/dbusbind.c (xd_read_message_1): Rework the code a bit
so that it has fewer tests. This pacifies GCC 12 which was
complaining incorrectly about dereferencing a null pointer.
* src/intervals.c (copy_properties): Remove an eassume that should
no longer be needed even to pacify older GCCs, due to ...
* src/intervals.h (split_interval_left): ... this addition of
ATTRIBUTE_RETURNS_NONNULL to pacify a GCC 12 warning about
dereferencing a null pointer.
* src/regex-emacs.c (EXTEND_BUFFER): Use negative values rather
than auxiliary booleans to indicate null pointers. This pacifies
GCC 12 false alarms about using uninitialized variables.
* src/xdisp.c (clear_position): New function.
(append_space_for_newline, extend_face_to_end_of_line):
Use it to work around false alarms from GCC 12.
(display_and_set_cursor): Add an UNINIT to pacify GCC 12.
* src/xterm.c (x_draw_glyphless_glyph_string_foreground):
Defend against hypothetical bad code elsewhere;
this also pacifies GCC 12.
(x_term_init): Use fixed-size auto array rather than alloca,
as the array is small; this also pacifies GCC 12.
This reverts commit 3cccf0a910.
This is a change with far-reaching effects on MS-Windows at the least,
where file-related APIs are shadowed to support transparent support
for UTF-8 encoded file names. Making such changes on a stable branch
for the benefit of a proprietary platform with a 13-year old OS is a
tail wagging the dog. Please don't do that without discussing first.
Use openat only on platforms with O_PATH.
This ports to OS X 10.9 and earlier.
Problem reported by Keith David Bershatsky in:
https://lists.gnu.org/r/emacs-devel/2022-04/msg00805.html
* lib-src/emacsclient.c (local_sockname): Use open, not openat.
* src/sysdep.c (sys_openat): New static function,
which uses openat only if O_PATH is defined.
(emacs_openat): Use it instead of openat.
(emacs_openat_noquit): Remove.
(emacs_open_noquit): Reimplement as per the old emacs_openat_noquit,
but use plain 'open'.
This is because mod-test.c shouldn’t use source code from lib,
but it does need to include <config.h>.
* lib-src/ebrowse.c, lib-src/emacsclient.c, lib-src/etags.c:
* lib-src/make-docfile.c, lib-src/movemail.c:
* lib-src/seccomp-filter.c, src/dynlib.h, src/lisp.h:
* src/syssignal.h, src/sysstdio.h, src/systhread.h, src/tparam.h:
Include <attribute.h>.
* src/conf_post.h: Do not include <attribute.h>.
All uses of attribute.h macros replaced with their _GL_ equivalents.
* lib-src/seccomp-filter.c (main): Allow 'pread64' and 'faccessat2'
system calls. Newer versions of glibc use these system call (starting
with commits 95c1056962a3f2297c94ce47f0eaf0c5b6563231 and
3d3ab573a5f3071992cbc4f57d50d1d29d55bde2, respectively).
11ea45c9e4 Fix UB in ebrowse
ba57b78064 Fix execute-extended-command-for-buffer in fundamental-mode
ef0c1d4c2c Add workaround to handle a problem with Enlightenment WM (...
* lib-src/ebrowse.c (matching_regexp): Avoid writing beyond the
limits of 'matching_regexp_buffer'. Patch by Jan Stranik
<jan@stranik.org>. (Bug#53333)
Copyright-paperwork-exempt: yes
* lib-src/be_resources.cc (main): Scale input bitmap
appropriately before setting it as the mini and large icons.
* src/Makefile.in (Emacs): Add dependency on
`lib-src/be-resources'.
While we don't need to put docstrings of .elc files into etc/DOC,
we still need to put those of `loaddefs.el` there since we don't have
a "dynamic docstring" feature for the non-compiled files and keeping
the actual docstrings in the heap would be prohibitive.
* src/Makefile.in ($(etc)/DOC): Scan `lisp/loaddefs.el` still.
* lib-src/make-docfile.c (scan_lisp_file): New function.
(scan_file): Use it.
(skip_white, read_lisp_symbol, search_lisp_doc_at_eol): New functions.
efb1c7ec37 ; * admin/notes/years: More known problems.
f475c5823a ; Copyright year fixes
2684def348 ; Run set-copyright from admin.el
2558b37b5e ; * admin/notes/years: Mention known problems with 'update...
Since the location of those files changes between build time and
installation time, this requires to tweak the file name used in those
(#$ . NNN) references during the dump so they don't hardcode the build
directory. We do it in the same way as was already done for those
same file names in `load-history`, except we convert them back to
absolute file names more lazily (i.e. when fetching the actual
docstring rather than at startup), which requires remembering the
`lisp-dir` computed at startup in the new `lisp-directory` variable.
* src/Makefile.in ($(etc)/DOC): Don't scan Lisp files any more.
* src/lread.c (Fload): Use relative file names for `load-file-name`
when preloading for the dump, like we already did for `current-load-list`.
(read_list): Don't zero-out dynamic docstring references during the
preload since they won't be filled later by Snarf-documentation any more.
(read1): Remove the hash-hack for doc references that were zeroed.
* lisp/startup.el (lisp-directory): New variable.
(command-line): Set it.
* src/doc.c (get_doc_string): Use `lisp-directory` for dynamic
docstring references using relative file names.
(syms_of_doc): Add `Qlisp_directory`.
* lib-src/make-docfile.c (scan_file): Don't handle `.el` or `.elc`
files any more.
(IS_SLASH): Remove macro, not used any more.
(skip_white, read_lisp_symbol, search_lisp_doc_at_eol)
(scan_lisp_file): Remove functions, not used any more.
* doc/lispref/loading.texi (Library Search): Mention `lisp-directory`.
