Paul Eggert [Mon, 26 Sep 2016 20:32:28 +0000 (13:32 -0700)]
Merge from origin/emacs-25
0ffc9ce Update admin/authors.el 0ad7410 Update Antinews in ELisp manual ea0f750 Fix comments on window height macros 0bbdeed Fix 'url-http-create-request' when cookies are used 0045998 Fix cross reference in frames.texi 1392894 ; * etc/DEBUG: Minor copyedits. 304a5c8 ; * etc/DEBUG: Improve documentation of getting control to GDB. 56bf7d7 Fix regexp-opt documentation (bug #17862) 803ad6f ; Fix documentation of seq-subseq ed4530d * lisp/emacs-lisp/gv.el (gv-ref): Fix example of PLACE in doc... 88ea396 ; Spelling fixes 17197d0 Fix tags-query-replace docstring 80a7f8b Clarify documentation of precision in format specs 88a5052 Improve and clarify documentation of subprocesses 89eb09f * etc/PROBLEMS: Mention gnutls-cli 3.5.3 (Bug#24247).
* lisp/textmodes/css-mode.el (css-class-list-function): New variable
holding the function to call for retrieving completions of class
names.
(css-id-list-function): New variable holding the function to call for
retrieving completions of IDs.
(css--foreign-completions): New function for retrieving completions
from other buffers.
(css--complete-selector): Support completing HTML class names and IDs
from other buffers in addition to completing HTML tags.
* lisp/textmodes/sgml-mode.el (html--buffer-classes-cache): New
variable holding a cache for `html-current-buffer-classes'.
(html--buffer-ids-cache): New variable holding a cache for
`html-current-buffer-ids'.
(html-current-buffer-classes): New function returning a list of class
names used in the current buffer.
(html-current-buffer-ids): New function returning a list of IDs used
in the current buffer.
(html-mode): Set `css-class-list-function' and `css-id-list-function'
to `html-current-buffer-classes' and `html-current-buffer-ids'
respectively.
Paul Eggert [Sat, 24 Sep 2016 09:35:13 +0000 (02:35 -0700)]
Improve integer overflow handling a bit
* src/charset.c (read_hex): Use INT_LEFT_SHIFT_OVERFLOW for clarity.
The machine code is the same on my platform.
* src/doprnt.c (doprnt):
* src/emacs-module.c (module_funcall):
* src/font.c (font_intern_prop):
* src/keyboard.c (Frecursion_depth):
* src/lread.c (read1):
Use WRAPV macros instead of checking overflow by hand.
* src/editfns.c (hi_time, time_arith, decode_time_components):
* src/emacs-module.c (Fmodule_load):
Simplify by using FIXNUM_OVERFLOW_P.
* src/emacs-module.c: Include intprops.h.
* src/xdisp.c (percent99): New function.
(decode_mode_spec): Use it to simplify overflow avoidance and
formatting of %p and %P.
Paul Eggert [Wed, 21 Sep 2016 22:17:20 +0000 (15:17 -0700)]
Port recent flexmember changes to NetBSD
The NetBSD library supports localtime_rz directly, and without
this change 'configure' omitted the flexmember test that is marked
as conditional on localtime_rz. Emacs now needs to use
flexmember.m4 even when localtime_rz works.
Problem reported by Thomas Klausner.
* admin/merge-gnulib (GNULIB_MODULES): Add flexmember.
* lib/gnulib.mk, m4/gnulib-comp.m4: Regenerate.
Paul Eggert [Tue, 20 Sep 2016 15:30:17 +0000 (08:30 -0700)]
Use flexmembers on IBM XL C for AIX
This removes a workaround where Emacs did not use flexible
array members when compiled with IBM XL C. Instead, avoid
the problem by making the aliasing issues more obvious to
this compiler.
* admin/merge-gnulib: Don’t remove m4/flexmember.m4.
* m4/flexmember.m4: Copy from gnulib.
* configure.ac (AC_C_FLEXIBLE_ARRAY_MEMBER): Remove workaround.
* src/alloc.c (allocate_string_data): Rephrase to avoid aliasing
problem that would otherwise mess up code generated for flexible
array members by IBM XL C for AIX, V12.1.
* src/conf_post.h (FLEXIBLE_ARRAY_MEMBER): Remove; now done
by gnulib code.
Michael Albinus [Mon, 19 Sep 2016 12:21:35 +0000 (14:21 +0200)]
Fix a problem with Tramp when the temporary dir isn't C:/Temp on w32
* lisp/net/tramp-sh.el (tramp-sh-handle-expand-file-name): Don't use
`tramp-drop-volume-letter' any longer, it causes collateral
damages. Reported by Chris Zheng <chriszheng99@gmail.com>.
Paul Eggert [Sat, 17 Sep 2016 22:06:24 +0000 (15:06 -0700)]
Port January __morecore changes to AIX 7.1
* src/vm-limit.c (__MALLOC_HOOK_VOLATILE, __morecore)
(__after_morecore_hook):
* src/ralloc.c (__morecore):
Declare if DOUG_LEA_MALLOC is not defined, not if HAVE_MALLOC_H is
not defined. <malloc.h> does not declare these in AIX 7.1.
Paul Eggert [Sat, 17 Sep 2016 18:05:07 +0000 (11:05 -0700)]
Define _GNU_SOURCE in files delaying config.h
Problem reported by Richard Copley in:
http://lists.gnu.org/archive/html/emacs-devel/2016-09/msg00440.html
* src/w32.c, src/w32notify.c, src/w32proc.c (_GNU_SOURCE):
Define early.
Eli Zaretskii [Sat, 17 Sep 2016 16:12:46 +0000 (19:12 +0300)]
Update the bidi test code and data
* test/manual/BidiCharacterTest.txt: Import the version from
Unicode Standard 9.0.0.
* test/manual/biditest.el (biditest-generate-testfile): Allow 'x'
in the 4th field of the test data.
Alan Mackenzie [Sat, 17 Sep 2016 12:43:54 +0000 (12:43 +0000)]
Improve accuracy of line/column numbers in byte compiler's warning messages.
* lisp/emacs-lisp/bytecomp.el (byte-compile-set-symbol-position): ensure new
value of byte-compile-last-position is not lower than old value.
(byte-compile-function-warn): call byte-compile-set-symbol-position.
Paul Eggert [Sat, 17 Sep 2016 01:01:39 +0000 (18:01 -0700)]
Define _GNU_SOURCE in unexmacosx.c
Problem reported by Bob Halley in:
http://lists.gnu.org/archive/html/emacs-devel/2016-09/msg00427.html
* src/unexmacosx.c (_GNU_SOURCE): Define if not already defined.
Alan Mackenzie [Fri, 16 Sep 2016 10:47:55 +0000 (10:47 +0000)]
C Mode: Fix mis-fontification of macro invocation as function declaration
This happened with a macro invocation which was followed by a statement block
in braces.
* lisp/progmodes/cc-engine.el (c-forward-decl-or-cast-1): CASE 11: distinguish
between contexts nil and top, being less permissive in the former case.
(c-just-after-func-arglist-p): call c-forward-decl-or-cast-1 with context top.
Paul Eggert [Thu, 15 Sep 2016 22:25:33 +0000 (15:25 -0700)]
Prefer now-standard int width macros
* src/data.c (ULL_WIDTH):
* src/lisp.h (EMACS_INT_WIDTH, BITS_PER_BITS_WORD):
* src/lread.c (read_integer):
* src/term.c (produce_glyphless_glyph):
* src/xterm.c (x_send_scroll_bar_event):
Use *_WIDTH macros instead of CHAR_BIT * sizeof.
* src/data.c (ULL_WIDTH): Rename from BITS_PER_ULL for consistency
with the *_WIDTH standard macros. All uses changed.
* src/gmalloc.c (INT_BIT): Remove. All uses replaced with INT_WIDTH.
* src/lisp.h (EMACS_INT_WIDTH): Rename from BITS_PER_EMACS_INT
for consistency with the *_WIDTH standard macros. All uses changed.
(BITS_PER_CHAR): Remove; all uses replaced by CHAR_BIT. This must
be the same as CHAR_WIDTH and avoids confusion with Elisp
char-width, which counts columns not bits.
(BITS_PER_SHORT): Remove; all uses replaced by SHRT_WIDTH.
(BITS_PER_LONG): Remove; all uses replaced by LONG_WIDTH.
* src/lread.c: Do not include limits.h since CHAR_BIT is no longer
used directly.
Paul Eggert [Thu, 15 Sep 2016 18:50:23 +0000 (11:50 -0700)]
Update from gnulib
This incorporates:
2016-09-15 stdint: support new _WIDTH macros
2016-09-15 limits-h: new module
2016-09-15 sys_types: avoid glibc 2.25 warnings about major()
2016-09-15 extensions: port to more ISO C TSes
2016-09-13 intprops: new macro TYPE_WIDTH
2016-09-13 extensions: port to recent ISO C TRs
* .gitignore: Add lib/limits.h.
* doc/misc/texinfo.tex, lib/ftoastr.h, lib/intprops.h:
* lib/stdint.in.h, m4/extensions.m4, m4/stdint.m4, m4/stdio_h.m4:
* m4/sys_types_h.m4: Copy from gnulib.
* lib/gnulib.mk, m4/gnulib-comp.m4: Regenerate.
* lib/limits.in.h, m4/limits-h.m4: New files, copie from gnulib.
* nt/gnulib.mk: Merge changes from lib/gnulib.mk.
Michael Albinus [Thu, 15 Sep 2016 07:24:07 +0000 (09:24 +0200)]
Rework fixing Bug#24432
* lisp/net/tramp.el (tramp-get-buffer): Set connection property
"process-buffer" in order to mark connection as active.
* lisp/net/tramp-cache.el (tramp-get-hash-table)
(tramp-set-connection-property)
(tramp-dump-connection-properties): Do not use "active" property.
(tramp-list-connections): Use "process-buffer" property.
* lisp/net/tramp-cmds.el (tramp-cleanup-connection): Flush process
properties prior deletion of process.
Eli Zaretskii [Wed, 14 Sep 2016 17:46:59 +0000 (20:46 +0300)]
Avoid compiler warning on MS-Windows due to clone_file
* src/fileio.c (clone_file): Condition on !WINDOWSNT, since the
call to it is not compiled in that build, and having it defined
causes compiler warnings.
Eli Zaretskii [Wed, 14 Sep 2016 17:41:04 +0000 (20:41 +0300)]
Avoid aborts in GC due to abort_on_gc
* src/lisp.h:
* src/print.c (Fprin1_to_string):
* src/eval.c (signal_or_quit):
* src/alloc.c (garbage_collect_1): Remove declarations, setting,
and testing the value of abort_on_gc. It is no longer needed, and
using it causes rare aborts in GC for no good reason. (Bug#23912)
Michael Albinus [Wed, 14 Sep 2016 09:49:56 +0000 (11:49 +0200)]
Fix Bug#24432
* lisp/net/tramp-cache.el (tramp-get-hash-table): Initialize a
connection as inactive.
(tramp-set-connection-property): Make the connection active.
(tramp-dump-connection-properties): Don't save "active" property.
(tramp-list-connections): List only active connections. (Bug#24432)
* lisp/net/sieve-manage.el (sieve-manage-ignore-starttls):
Rename from sieve-manage-ignore-broken-tls. If it is set,
sieve-manage will never use STARTTLS even if the server says
it is capable (but may be broken).
dired-jump: Expand file-name before dired-goto-file call
Command dired-goto-file requires its argument to be an absolute
file name. Interactively FILE-NAME is read with read-file-name,
which could return an abbreviated file name (Bug#24409).
* lisp/dired-x.el (dired-jump): Use expand-file-name on FILE-NAME.
Clarify in doc string the meaning of arg FILE-NAME.
sieve-manage.el: Allow user to avoid STARTTLS capability test (bug#24422)
* lisp/net/sieve-manage.el
(sieve-manage-ignore-broken-tls): New user option.
(sieve-manage-open-server):
Don't test STARTTLS capability if the option is set (bug#24422).
Fix compiler thinking width and height may be unitialised in frame.c
This fixes the following warning:
frame.c: In function ‘x_set_frame_parameters’:
frame.c:3329:25: error: ‘width’ may be used uninitialized in
this function [-Werror=maybe-uninitialized]
adjust_frame_size (f, width_change ? width : -1,
^
* src/frame.c (x_set_frame_parameters): Drop width_changed and
height_changed variables in favour of storing that information in
width and height variables.
Fix compiler thinking tmpdir may be unitialised in emacsclient
This fixes the following warning:
emacsclient.c: In function ‘set_local_socket’:
/usr/include/x86_64-linux-gnu/bits/string3.h:111:3: error:
‘tmpdir’ may be used uninitialized in this function
[-Werror=maybe-uninitialized]
return __builtin___stpcpy_chk (__dest, __src, __bos (__dest));
^
emacsclient.c:1197:17: note: ‘tmpdir’ was declared here
const char *tmpdir;
* lib-src/emacsclient.c (set_local_socket): Get rid of use_tmpdir
variable and instead use tmpdir being non-NULL as sign that it should
be used.
image-dired: Signal an error before calling a missing executable
Reverts commit ca473907
Add a defun to check if an executable exists, and call it
on each function using an external program: when the executable
is not available signal an error.
See discussion on:
https://lists.gnu.org/archive/html/emacs-devel/2016-09/msg00135.html
* lisp/image-dired.el (image-dired--check-executable-exists): New defun.
Throw and error when the executable arg is missing.
(image-dired-display-image, image-dired-rotate-thumbnail)
(image-dired-rotate-original, image-dired-set-exif-data)
(image-dired-get-exif-data):
Use it.
Refactor common code in {upcase,downcase,capitalize}-word functions
* src/casefiddle.c (operate_on_word): Removed in favour of…
(casify_word) …new function which does what operate_on_word did plus
what all of the common code from *-word functions.
(upcase-word, downcase-word, capitalize-word): Move code common between
those functions (pretty much the whole body of those functions) into
casify_word and use that instead of now deleted operate_on_word.
Alan Mackenzie [Sun, 11 Sep 2016 21:09:08 +0000 (21:09 +0000)]
Correctly fontify C++ direct initializations with parens inside functions
Or, more clearly, when something looks like a function declaration and it's
inside a function, fontify it as a direct initialization.
For this purpose, introduce a "brace stack" for each buffer, where an entry on
the brace stack states how deeply nested a particular position is inside
braces inside a "top level", which includes classes and namespaces.
Also introduce a new "context", "top", with which c-font-lock-declarations
signals to c-forward-decl-or-cast-1 that point is at the top level.
* lisp/progmodes/cc-langs.el (c-get-state-before-change-functions): add
c-truncate-bs-cache.
(c-flat-decl-block-kwds, c-brace-stack-thing-key, c-brace-stack-no-semi-key)
(c-type-decl-operator-prefix-key): new language constants/variables.
* lisp/progmodes/cc-engine.el (c-bs-interval, c-bs-cache, c-bs-cache-limit)
(c-bs-prev-pos, c-bs-prev-stack): New mostly local variables for the brace
stack cache.
(c-init-bs-cache, c-truncate-bs-cache, c-truncate-bs-cache, c-brace-stack-at)
(c-bs-at-toplevel-p): New functions which manipulate the brace stack (cache).
(c-find-decl-prefix-search): Keep track of whether we're at top level.
(c-find-decl-spots): New local variable cfd-top-level which records what it
says. On calling cfd-fun, pass cfd-top-level as an additional argument.
(c-forward-declarator): Add new element DECORATED to the result list. Set it
to non-nil when a match for c-type-decl-operator-prefix-key is found.
(c-forward-decl-or-cast-1): Handle the newly introduced context "top".
Introduce "CASE 9.5", which recognizes direct initializations.
* lisp/progmodes/cc-fonts.el (c-font-lock-complex-decl-prepare)
(c-font-lock-enum-tail, c-font-lock-cut-off-declarators)
(c-font-lock-enclosing-decls, c-simple-decl-matchers, c-basic-matchers-after):
Add appropriate `not-top' argument to calls to c-font-lock-declarators.
(c-font-lock-declarators): Additional parameter `not-top'. Use not-top to
participate in the decision whether to fontify an identifier as a function or
a variable.
(c-font-lock-declarations): The internal lambda function takes an additional
argument `toplev' from c-find-decl-spots, which it uses in determining the
"context" of a declaration. Add appropriate `not-top' argument to calls to
c-font-lock-declarators.
(c-font-lock-objc-methods): Add extra parameter to internal lambda function,
like for c-font-lock-declarators.
* lisp/progmodes/cc-mode.el (c-basic-common-init): Initialize the brace stack
cache.
* src/window.h (WINDOW_MODE_LINE_HEIGHT, WINDOW_HEADER_LINE_HEIGHT): Fix
confusing claim that "height is in pixels and in lines"; in fact it's in
pixels.
For errors, use ‘byte-compile-report-error’ instead so that the error
is registered and causes compilation to fail (Bug#24359).
For warnings, use ‘byte-compile-warn’ instead so that
‘byte-compile-error-on-warn’ is honored (Bug#24360).
* lisp/emacs-lisp/macroexp.el (macroexp--funcall-if-compiled)
(macroexp--warn-and-return): Use ‘byte-compile-warn’ instead of
‘byte-compile-log-warning’.
* lisp/emacs-lisp/bytecomp.el (byte-compile-form, byte-compile-unfold-bcf)
(byte-compile-setq, byte-compile-funcall): Use
‘byte-compile-report-error’ instead of ‘byte-compile-log-warning’.
(byte-compile-log-warning): Convert comment to documentation
string. Explain that the function shouldn’t be called directly.
(byte-compile-report-error): Add optional FILL argument.
* lisp/emacs-lisp/cconv.el (cconv-convert, cconv--analyze-use)
(cconv--analyze-function, cconv-analyze-form): Use
‘byte-compile-warn’ instead of ‘byte-compile-log-warning’.
* lisp/emacs-lisp/byte-opt.el (byte-compile-inline-expand): Use
‘byte-compile-warn’ instead of ‘byte-compile-log-warning’.
* lisp/subr.el (add-to-list): Use ‘byte-compile-report-error’ instead
of ‘byte-compile-log-warning’.
(do-after-load-evaluation): Use ‘byte-compile-warn’ instead of
‘byte-compile-log-warning’.
Paul Eggert [Sat, 10 Sep 2016 19:51:27 +0000 (12:51 -0700)]
copy-file now uses GNU/Linux file cloning
From a suggestion by Kieran Colford (see Bug#23904).
* configure.ac: Check for linux/fs.h.
* src/fileio.c [HAVE_LINUX_FS_H]: Include sys/ioctl.h and linux/fs.h.
(clone_file): New function.
(Fcopy_file): Use it.
Fix 'url-http-create-request' when cookies are used
* lisp/url/url-http.el (url-http-create-request): Make sure the
cookie headers are a unibyte string. For the details, see
http://lists.gnu.org/archive/html/emacs-devel/2016-09/msg00202.html.
It is useful to be able to call `isearch-done' unconditionally to
ensure a non-isearching state.
* lisp/isearch.el (isearch-done): Check that `isearch--current-buffer'
is a live buffer before using it (Bug #21091).
* test/lisp/isearch-tests.el (isearch--test-done): Test it.
* lisp/emacs-lisp/ring.el (ring-elements): Don't use the RESULT
argument of `dotimes' when the iteration variable isn't referred by
it.
(ring-member): Don't pass nil as the RESULT argument of `dotimes'
since it's the default.
Split regex character class test into smaller chunks
Having one test for all character classes it is not always trivial to
determine which class is failing. This happens when failure is caused
by ‘(should (equal (point) (point-max)))’ not being met.
With per-character class tests, it is immidiatelly obvious which test
causes issues plus tests for all classes are run even if some of them
fail.
Don’t allocate char-table’s extra slots in regexp-out-charset
* lisp/emacs-lisp/regexp-opt.el (regexp-opt-charset): Do not use
'case-table as charmap char-table’s property. The function has nothing
to do with casing and in addition using 'case-table causes unnecessary
extra slots to be allocated which ‘regexp-opt-charset’ does not use.
RE_CHAR_TO_MULTIBYTE(c) yields c for ASCII characters and a byte8
character for c ≥ 0x80. Furthermore, CHAR_BYTE8_P(c) is true only
for byte8 characters. This means that
c = RE_CHAR_TO_MULTIBYTE (ch);
if (! CHAR_BYTE8_P (c) && re_iswctype (c, cc))
is equivalent to:
c = c;
if (! false && re_iswctype (c, cc))
for 0 ⪬ c < 0x80, and
c = BYTE8_TO_CHAR (c);
if (! true && re_iswctype (c, cc))
for 0x80 ⪬ c < 0x100. In other words, the loop never executes for
c ≥ 0x80 and RE_CHAR_TO_MULTIBYTE call is unnecessary for c < 0x80.
* src/regex.c (regex_compile): Simplyfy a for loop by eliminating
dead iterations and unnecessary macro calls.
decimalnump was used in regex.c only in ISALNUM macro which ored it with
alphabeticp. Because both of those functions require Unicode general
category lookup, this resulted in unnecessary lookups (if alphabeticp
return false decimalp had to perform another lookup). Drop decimalnump
in favour of alphanumericp which combines decimelnump with alphabeticp.
* src/character.c (decimalnump): Remove in favour of…
(alphanumericp): …new function.
STRING_CHAR does not unify characters; update documentation
* src/character.h (STRING_CHAR): Update doc.
* src/buffer.h (FETCH_MULTIBYTE_CHAR): Update doc. While at it, change
the function to use BYTE_POS_ADDR instead of open-coding it.
Paul Eggert [Thu, 8 Sep 2016 01:08:45 +0000 (18:08 -0700)]
Port flexible array members to GCC + valgrind
These changes are needed to conform to the C standard's rule for
allocating structs containing flexible array members. C11 says
that malloc (offsetof (struct s, m) + n) does not suffice to
allocate a struct with an n-byte tail; instead, malloc’s arg
should be rounded up to the nearest multiple of alignof (struct s).
Although this is arguably a defect in C11, gcc -O2 + valgrind
sometimes complains when this rule is violated, and when debugging
it’s better to keep valgrind happy.
For details please see the thread containing the message at:
https://gcc.gnu.org/ml/gcc-patches/2016-09/msg00416.html
* lib-src/ebrowse.c, src/alloc.c, src/image.c, src/process.c:
Include flexmember.h.
* lib-src/ebrowse.c (add_sym, add_member, make_namespace)
(register_namespace_alias):
* src/alloc.c (SDATA_SIZE, allocate_string_data):
* src/image.c (xpm_cache_color, imagemagick_create_cache):
* src/process.c (Fmake_network_process):
Use FLEXSIZEOF instead of offsetof and addition.
* src/alloc.c (SDATA_SIZE, vector_alignment):
Use FLEXALIGNOF instead of sizeof (ptrdiff_t).
* src/lisp.h (ALIGNOF_STRUCT_LISP_VECTOR):
Remove, as alloc.c can now calculate this on its own.
Paul Eggert [Thu, 8 Sep 2016 00:04:49 +0000 (17:04 -0700)]
Update from gnulib
This incorporates:
2016-09-07 flexmember: new macro FLEXALIGNOF
2016-09-07 flexmember: port better to GCC + valgrind
2016-08-18 Port modules to use getprogname explicitly
2016-09-02 manywarnings: add -fno-common
* admin/merge-gnulib (GNULIB_TOOL_FLAGS): Don’t avoid flexmember,
since time_rz now uses part of it. Instead, remove m4/flexmember.m4.
* configure.ac (AC_C_FLEXIBLE_ARRAY_MEMBER): Define away,
since Emacs assumes C99 and therefore removes m4/flexmember.m4.
* lib/euidaccess.c, lib/group-member.c, lib/time_rz.c:
* m4/manywarnings.m4: Copy from gnulib.
* lib/flexmember.h: New file, from gnulib.
* lib/gnulib.mk, m4/gnulib-comp.m4: Regenerate.
Noam Postavsky [Sun, 21 Aug 2016 14:51:38 +0000 (10:51 -0400)]
Don't --load directories
* lisp/startup.el (command-line-1): Only pass expanded FILENAME argument
of --load when it refers to a normal file, since `load' doesn't handle
directories (Bug #16406).
* lisp/calendar/icalendar.el (icalendar--read-element): Avoid a regex
stack overflow by not using regex to extract values from calendar
events. (Bug#24315)
projects / emacs.git / log