Andrew Hyatt [Thu, 4 Feb 2016 05:13:36 +0000 (00:13 -0500)]
Remove packages obsoleted before Emacs 24.
In accordance with the policy discussed in the emacs-devel list,
packages that have been obsoleted for a full major release cycle are up
for deletion.
This removes almost all packages that are now eligible for deletion,
with the exception of "cl-compat", which seems it is likely to still be
used, and "optional", which offers some functionality that doesn't have
a replacement yet.
Paul Eggert [Wed, 3 Feb 2016 21:21:10 +0000 (13:21 -0800)]
Mention context when resume from emergency escape
That way, if the user has been doing something else for a while,
they are reminded of the situation when restarting Emacs,
and are more likely to understand the two questions.
* doc/emacs/trouble.texi (Emergency Escape): Document this.
* src/keyboard.c (handle_interrupt): Implement this.
Paul Eggert [Wed, 3 Feb 2016 08:37:44 +0000 (00:37 -0800)]
Port aligned_alloc decl to Cygwin.
Problem reported by Ken Brown (Bug#22522#38).
* configure.ac (aligned_alloc): Check for decl too.
* src/lisp.h (aligned_alloc): Declare if not already declared.
Paul Eggert [Wed, 3 Feb 2016 07:11:36 +0000 (23:11 -0800)]
autogen.sh now configures git only on request
* autogen.sh (do_autoconf, do_git): New vars.
Support new arguments --help, all, autoconf, git.
By default, just do autoconf-related configuration, not git.
Prefer 'echo' to 'cat <<EOF ...', as this tends to avoid temp files.
If GNU cp is available, use it to backup .git/config before
changing it. When configuring git, chatter about what is being
done, and configure git to check hashes. Avoid some duplicate
file name specification when creating git hooks.
* GNUmakefile (ALL_IF_GIT): New macro.
(configure): Use it.
* INSTALL.REPO: Suggest './autogen.sh all'.
Paul Eggert [Sun, 31 Jan 2016 17:50:07 +0000 (09:50 -0800)]
Port new hybrid malloc to FreeBSD
Problem reported by Wolfgang Jenkner in: http://bugs.gnu.org/22086#118
* src/gmalloc.c (__malloc_initialize_hook, __after_morecore_hook)
(__morecore) [HYBRID_MALLOC]: Define in this case too.
Michael Albinus [Sun, 31 Jan 2016 12:50:23 +0000 (13:50 +0100)]
Merge changes from Tramp repository
* doc/misc/Makefile.in (${buildinfodir}/tramp.info tramp.html):
No EXTRA_OPTS needed.
* doc/misc/tramp.texi: Merge changes from Emacsemacs-25
branch, especially for @trampfn{}.
(Top): Move @ifnottex down.
(History): XEmacs support has been removed.
(GVFS based methods, Remote processes): Do not use emacsgvfs flag.
(Auto-save and Backup): Use both syntax versions.
(File name Syntax): Remark on IPv6 adresses is valid for
unified syntax only.
* doc/misc/trampver.texi: Do not set emacsgvfs flag.
Paul Eggert [Sat, 30 Jan 2016 22:20:57 +0000 (14:20 -0800)]
Shrink static heap a bit
* src/sheap.h: Include lisp.h, for Lisp_Object.
(STATIC_HEAP_SIZE): Now an enum constant, not a macro.
Make it 2 MiB * sizeof (Lisp_Object), which is a bit more
conservative than the old value.
(Bug#22086)
Paul Eggert [Sat, 30 Jan 2016 22:20:57 +0000 (14:20 -0800)]
Fix extern symbols defined and not used
* src/alloc.c: Always include <signal.h>.
(malloc_warning) [!SIGDANGER && (SYSTEM_MALLOC || HYBRID_MALLOC)]:
Do not define; unused.
* src/emacs.c, src/lisp.h (might_dump) [!DOUG_LEA_MALLOC]: Now static.
* src/gmalloc.c (gdefault_morecore): Rename from __default_morecore,
to avoid collision with glibc. Now static. All uses changed.
* src/lastfile.c (my_edata): Define only if
((!defined SYSTEM_MALLOC && !defined HYBRID_MALLOC && !defined
WINDOWSNT) \ || defined CYGWIN || defined DARWIN_OS).
(Bug#22086)
Paul Eggert [Sat, 30 Jan 2016 22:20:57 +0000 (14:20 -0800)]
Build lib/e-*.o only on platforms that need it
* configure.ac (hybrid malloc): Simplify configuration.
(SHEAP_OBJ): Remove; no longer needed.
(HYBRID_MALLOC): New var. Subst it.
(HYBRID_MALLOC_LIB): New Automake conditional.
* lib/Makefile.am (noinst_LIBRARIES): Add libegnu.a only if
HYBRID_MALLOC_LIB.
(libegnu_a_CPPFLAGS): Omit AM_CPPFLAGS; not needed.
(MOSTLYCLEANFILES): Add libegnu.a.
* src/Makefile.in (SHEAP_OBJ): Remove.
(HYBRID_MALLOC): New macro.
(base_obj): Use it to conditionally add sheap.o.
(LIBEGNU_ARCHIVE): New macro.
($(LIBEGNU_ARCHIVE)): New rule, replacing $(lib)/libegnu.a.
All uses of the latter replaced by the former.
* src/alloc.c (USE_ALIGNED_ALLOC): Simplify configuration.
Correct misspelling ALIGNED_ALLOC to HAVE_ALIGNED_ALLOC.
* src/gmalloc.c: Update comment.
* src/lisp.h (aligned_alloc)
[!DOUG_LEA_MALLOC && !HYBRID_MALLOC && !SYSTEM_MALLOC]:
New decl.
(Bug#22086)
Paul Eggert [Sat, 30 Jan 2016 22:20:57 +0000 (14:20 -0800)]
Include <malloc.h> when advisable
This should help insulate us better from future glibc changes.
It is good hygiene to include .h files for APIs that Emacs uses.
Fix type clashes between Emacs and GNU <malloc.h> (Bug#22086).
* configure.ac: Check for malloc.h.
* src/alloc.c: Include <malloc.h> depending on HAVE_MALLOC_H,
not on DOUG_LEA_MALLOC.
* src/emacs.c, src/gmalloc.c (malloc_enable_thread):
Remove decl (now in lisp.h).
* src/gmalloc.c: Include stddef.h earlier, for ptrdiff_t.
[emacs]: Include lisp.h.
[HAVE_MALLOC_H]: Include <malloc.h>.
(__MALLOC_HOOK_VOLATILE): New macro, if not already defined.
(__after_morecore_hook, __malloc_initialize_hook, __morecore)
(__default_morecore):
[!HAVE_MALLOC_H]: New decls near non-inclusion of <malloc.h>.
(calloc): Make it clear that the macro should not be used.
Remove unused decl.
(malloc_info): New macro, to avoid clash with glibc <malloc.h>.
(__morecore, __default_morecore, __after_morecore_hook)
(__malloc_extra_blocks, __malloc_initialize_hook, __free_hook)
(__malloc_hook, __realloc_hook, __memalign_hook, memory_warnings):
Remove later decls.
(gmalloc_hook, gfree_hook, grealloc_hook):
Rename from __malloc_hook, __free_hook, __realloc_hook to
avoid type collision with glibc <malloc.h>. All uses changed.
(gmalloc_hook):
(__malloc_extra_blocks) [DOUG_LEA_MALLOC||HYBRID_MALLOC||SYSTEM_MALLOC]:
Now static.
(gmalloc_hook, __malloc_extra_blocks): Define even if [!HYBRID_MALLOC].
(__malloc_initialize_hook, __after_morecore_hook):
Declare with types compatible with glibc.
(__memalign_hook, hybrid_calloc) [HYBRID_MALLOC]:
Remove. All uses removed.
* src/lisp.h (__malloc_extra_blocks, malloc_enable_thread): New decls.
* src/ralloc.c, src/vm-limit.c:
Simplify includes and include <malloc.h> if available.
Paul Eggert [Sat, 30 Jan 2016 22:20:57 +0000 (14:20 -0800)]
* src/alloc.c: Include "sheap.h".
(alloc_unexec_pre, alloc_unexec_post) [HYBRID_MALLOC]:
Set and clear bss_sbrk_did_unexec, on all platforms not just Cygwin.
* src/lisp.h (alloc_unexec_pre, alloc_unexec_post) [!DOUG_LEA_MALLOC]:
Declare unconditionally.
* src/unexcw.c, src/unexelf.c (bss_sbrk_did_unexec): Remove decl.
(unexec): Don’t set or clear bss_sbrk_did_unexec;
the caller now does this.
(Bug#22086)
Paul Eggert [Sat, 30 Jan 2016 22:20:57 +0000 (14:20 -0800)]
Pacify --enable-gcc-warnings when HYBRID_MALLOC
* src/buffer.c (init_buffer):
* src/emacs.c (main):
* src/xsmfns.c (smc_save_yourself_CB, x_session_initialize):
Use emacs_get_current_dir_name, not get_current_dir_name.
* src/conf_post.h (aligned_alloc) [HYBRID_MALLOC && emacs]: New macro.
(HYBRID_GET_CURRENT_DIR_NAME, get_current_dir_name): Remove.
* src/emacs.c: Include "sheap.h".
(report_sheap_usage): Remove decl.
(Fdump_emacs) [HYBRID_MALLOC]: Report usage directly.
Don't assume ptrdiff_t can be printed as int.
* src/gmalloc.c [HYBRID_MALLOC]:
Include "sheap.h" rather than declaring its contents by hand.
(get_current_dir_name, gget_current_dir_name)
(hybrid_get_current_dir_name): Remove.
(emacs_abort): Remove duplicate decl.
(aligned_alloc): Undef, like malloc etc.
(ALLOCATED_BEFORE_DUMPING): Now a static function, not a macro.
Make it a bit more efficient.
(malloc_find_object_address): Remove unused decl.
(enum mcheck_status, mcheck, mprobe, mtrace, muntrace, struct mstats)
(mstats, memory_warnings): Declare only if GC_MCHECK.
* src/lisp.h (emacs_get_current_dir_name):
New decl, replacing get_current_dir_name.
* src/sheap.c: Include sheap.h first.
(STATIC_HEAP_SIZE): Remove; now in sheap.h.
(debug_sheap): Now static.
(bss_sbrk_buffer_end): Remove; no longer used.
(bss_sbrk_ptr): Now static and private.
(bss_sbrk_did_unexec): Now bool.
(BLOCKSIZE): Remove, to avoid GCC warning about its not being used.
(bss_sbrk): Don't treat request_size 0 as special, since the code
works without this being a special case.
Avoid overflow if request size exceeds INT_MAX.
(report_sheap_usage): Remove; now done in emacs.c.
* src/sheap.h: New file.
* src/sysdep.c (get_current_dir_name): Remove macro.
Include "sheap.h".
(emacs_get_current_dir_name): Rename function from
get_current_dir_name. Handle HYBRID_MALLOC here;
this is simpler.
(Bug#22086)
Paul Eggert [Sat, 30 Jan 2016 22:20:56 +0000 (14:20 -0800)]
Report static heap usage on non-Cygwin, too
* src/emacs.c (Fdump_emacs) [HYBRID_MALLOC]: Report sheap usage here ...
* src/unexcw.c (unexec): ... instead of here, since sheap can be used
on platforms other than Cygwin (Bug#22086).
Wolfgang Jenkner [Sat, 30 Jan 2016 22:20:56 +0000 (14:20 -0800)]
Add musl patch to support HYBRID_MALLOC on elf systems
* src/gmalloc.c: Adjust for HYBRID_MALLOC in the non CYGWIN case.
(__default_morecore): Here, in particular.
* configure.ac: Define HYBRID_MALLOC when unexelf.o is used.
New variable SHEAP_OBJ.
* src/Makefile.in: Use it.
(Bug#22086)
Wolfgang Jenkner [Sat, 30 Jan 2016 22:20:56 +0000 (14:20 -0800)]
Link temacs with gnulib compiled with -Demacs
This is done to support HYBRID_MALLOC, since some static variables
(e.g., last_environ in putenv.c) hold pointers to memory malloced
before dumping (Bug#22086).
* lib/Makefile.am: Add incantation to install libegnu.a.
* src/Makefile.in ($(lib)/libgnu.a): Replace with libegnu.a
(temacs$(EXEEXT)): Use it.
Wolfgang Jenkner [Sat, 30 Jan 2016 22:20:55 +0000 (14:20 -0800)]
Internal linkage for gmalloc etc. if HYBRID_MALLOC
This avoids clashes with symbols if the after-dump malloc is
derived from Doug Lea's implementation (Bug#22086).
* src/gmalloc.c (emacs_abort, __morecore, __default_morecore):
Move declarations up. For HYBRID_MALLOC, turn all `extern'
declarations below to `static' ones.
(aligned_alloc): Declare for !MSDOS as well.
(heapsize, _fraghead): Move resp. copy declaration downwards.
For HYBRID_MALLOC, conditionalize out the other definitions,
since the previous `static' declarations double as tentative
definitions, anyway.
(_malloc, _free, _realloc, __free_hook, _aligned_blocks)
(__realloc_hook, __memalign_hook): Conditionalize out.
(cfree, memalign, valloc): Ditto.
Paul Eggert [Sat, 30 Jan 2016 19:43:27 +0000 (11:43 -0800)]
; Merge from origin/emacs-25
The following commits were skipped:
ad879b7 Backport:fix previous change of src/ftfont.c (ftfont_shape_by_flt) 4a3db0f support rendering of wider range of combinging characters by ftfont backend
Paul Eggert [Sat, 30 Jan 2016 19:28:37 +0000 (11:28 -0800)]
Merge from origin/emacs-25
3f481ad Rename xref-query-replace to xref-query-replace-in-results 62f4ed4 Update cl-defgeneric and cl-defmethod docstrings 2111e0e Comment out next-error-function integration in xref 4e11ad3 Correct a use of "which" in intro.texi a1865bc Distinguish the two meanings of Java's keyword "default". Fixes bug #22358. 76045f7 Don't operate on menu bar of nonexistent frame c32f3bc Unbreak the GNUstep build.
lu4nx [Sat, 30 Jan 2016 12:56:43 +0000 (14:56 +0200)]
Support Go language in 'etags'
* lib-src/etags.c <Ruby_help>: Fix documentation of Ruby tags.
<Go_help>: New help.
<Go_suffixes>: New variable.
(Go_functions): New function.
<lang_names>: Add entry for Go. (Bug#22370)
* doc/emacs/maintaining.texi (Tag Syntax): Document Go support.
* doc/man/etags.1: Mention Go support.
* etc/NEWS: Mention Go support.
* test/etags/go-src/test.go:
* test/etags/go-src/test1.go: New test files.
* test/etags/Makefile (GOSRC): New variable.
(SRCS): Add $(GOSRC).
* test/etags/ETAGS.good_1:
* test/etags/ETAGS.good_2:
* test/etags/ETAGS.good_3:
* test/etags/ETAGS.good_4:
* test/etags/ETAGS.good_5:
* test/etags/ETAGS.good_6:
* test/etags/CTAGS.good: Adapt to addition of Go tests.
Eli Zaretskii [Sat, 30 Jan 2016 12:16:36 +0000 (14:16 +0200)]
Improve Ruby support in 'etags'
* lib-src/etags.c (Ruby_functions): Tag constants. Don't tag
singleton classes. Remove class qualifiers from tags generated
for method and constant names. (Bug#22241)
* doc/emacs/maintaining.texi (Tag Syntax): Mention that constants
are tagged by etags in Ruby.
* etc/NEWS: Mention that constants are tagged by etags in Ruby.
* test/etags/ruby-src/test1.ruby: Add more tests.
* test/etags/ETAGS.good_1:
* test/etags/ETAGS.good_2:
* test/etags/ETAGS.good_3:
* test/etags/ETAGS.good_4:
* test/etags/ETAGS.good_5:
* test/etags/ETAGS.good_6:
* test/etags/CTAGS.good: Adapt to the changes in etags and in Ruby
tests.
Eli Zaretskii [Sat, 30 Jan 2016 10:18:43 +0000 (12:18 +0200)]
Adjust etags test results to changes in copyright years
* test/etags/CTAGS.good:
* test/etags/ETAGS.good_1:
* test/etags/ETAGS.good_2:
* test/etags/ETAGS.good_3:
* test/etags/ETAGS.good_4:
* test/etags/ETAGS.good_5:
* test/etags/ETAGS.good_6: Adjust to shift in characters and
in line numbers.
Eli Zaretskii [Sat, 30 Jan 2016 08:04:13 +0000 (10:04 +0200)]
Document xwidget commands and functions
* doc/lispref/display.texi (Xwidgets): New section, describes some
of the xwidget primitives.
* doc/lispref/display.texi (Display): Update the chapter menu.
* doc/emacs/misc.texi (Embedded WebKit Widgets): New section.
* doc/emacs/emacs.texi (Top): Update the master menu to include
the xwidget node.
Dmitry Gutov [Sat, 30 Jan 2016 04:21:31 +0000 (07:21 +0300)]
Improve project-find-file yet again!
* lisp/progmodes/project.el (project--completing-read-strict):
New function.
(project-find-file-in): Use it.
(project-file-completion-table): Move the default
implementation inside the cl-defgeneric form.
(http://lists.gnu.org/archive/html/emacs-devel/2016-01/msg01720.html)
Stephen Leake [Fri, 29 Jan 2016 23:43:26 +0000 (17:43 -0600)]
Improve project-find-file
* lisp/progmodes/project.el (project-file-completion-table): New.
(project-find-file, project-or-external-find-file): Default to filename
at point.
(project-file-completion-table): New, split out from
project--find-file-in.
(project-find-file-in): Renamed from project--find-file-in, use
project-file-completion-table.
* lisp/progmodes/xref.el (ede-minor-mode): New declaration.
(xref--find-ignores-arguments): Add doc string.
* lisp/ses.el (ses-localvars): rename variable
`ses--renamed-cell-symb-list' into `ses--in-killing-named-cell-list'
and adjust the comment about it.
(ses-plist-delq): new defun.
(ses--ses-buffer-list): new defvar.
(ses--unbind-cell-name): new defun.
(ses-relocate-symbol): Do not relocate symbol when it is a named cell.
(ses-relocate-formula): Undo change of
2011-12-27T19:30:39Z!vincentb1@users.sourceforge.net that was
preventing relocation for named cell --- now doing this is delegated
to function `ses-relocate-symbol'.
(ses-relocate-range): In docstring, undo change of
2016-01-03T07:31:52Z!johnw@newartisans.com, `ses-range' must remain
lower case as it is not a variable.
(ses-relocate-all): Cell name relocation : 1) check that cell is a
renamed cell by testing `ses-cell' property to :ses-named, rather than
comparing name to corresponding standard name. Set rowcol of renamed
cell into the hashmap --- `ses-cell' property must not be used for
that as the same name can be used for different locations in different
SES sheets ; 2) use `local-variable-if-set-p' rather than `boundp' and
`local-variable-p' to check if cell name is already in use in this
sheet or needs initialisation.
(ses-relocate-all): Cell value relocation : 1) like for name
relocation use the `ses-cell' property rather than comparing actual
name to corresponding standard name. 2) Correct bug introduced in
2011-12-27T19:30:39Z!vincentb1@users.sourceforge.net, as the test was
made the other way round than the intention --- ie value relocation
was disabled for standard cell, not for renamed cell as was the
intention.
(ses-relocate-all): Add loop for unbinding deleted renamed cells
names.
(ses-killbuffer-hook): new defun.
(ses-mode): Add the ses--ses-buffer-list maintenance mechanism ---
kill buffer hook, plus pushing current buffer if new in list.
(ses-delete-row, ses-delete-column): Collect deleted renamed cells
into `ses--in-killing-named-cell-list'.
(ses-rename-cell): Remove update of variable
`ses--renamed-cell-symb-list', this variable is renamed to
`ses--in-killing-named-cell-list', and its setting is done in
functions `ses-delete-row' and , `ses-delete-column' now.
(ses-rename-cell): Make cell new name a buffer local variable.
(ses-rename-cell): Change correction of
2015-12-30T23:10:37Z!vincentb1@users.sourceforge.net concerning
computation of the range over which `cursor-intangible' property was
to be updated. This correction was ok for non spilling cells, but not
for cells spilling over following blank cells. Simply use
`next-single-property-change' rather than computing the end column
from column widths.
Eli Zaretskii [Fri, 29 Jan 2016 09:40:31 +0000 (11:40 +0200)]
Minor improvements to 'pcase' documentation
* doc/lispref/control.texi (Pattern matching case statement):
Improve the documentation of 'pcase' per comments. See two
discussion threads on emacs-devel@gnu.org for the details:
http://lists.gnu.org/archive/html/emacs-devel/2016-01/msg01335.html
http://lists.gnu.org/archive/html/emacs-devel/2016-01/msg01336.html.
Paul Eggert [Wed, 27 Jan 2016 07:00:10 +0000 (23:00 -0800)]
malloc.h hygiene
This attempts to future-proof Emacs a bit against possible glibc
changes, by having Emacs use <malloc.h> declarations rather than
coding them up by hand. Problem noted by Florian Weimer in:
https://sourceware.org/ml/libc-alpha/2016-01/msg00777.html
Implement this mainly by moving malloc.h-related functions from
emacs.c (which does not include <malloc.h>) to alloc.c (which does).
* src/alloc.c (my_heap_start) [DOUG_LEA_MALLOC || GNU_LINUX]:
New function.
The remaining changes to this file apply only if DOUG_LEA_MALLOC.
(alloc_unexec_pre, alloc_unexec_post): New functions.
(malloc_initialize_hook): Use my_heap_start and alloc_unexec_post.
(__MALLOC_HOOK_VOLATILE): New macro, if not already defined.
(__malloc_initialize_hook): Use it.
(malloc_state_ptr, malloc_initialize_hook, __malloc_initialize_hook):
Move here from ...
* src/emacs.c: ... here.
(malloc_get_state, malloc_set_state): Remove extern decls.
(my_heap_start) [DOUG_LEA_MALLOC || GNU_LINUX]: Remove static var.
All uses changed to similarly-named new function.
(Fdump_emacs): Use new functions alloc_unexec_pre, alloc_unexec_post.
* src/lisp.h (my_heap_start, alloc_unexec_pre, alloc_unexec_post):
New decls.
K. Handa [Sat, 2 Jan 2016 07:36:21 +0000 (16:36 +0900)]
support rendering of wider range of combinging characters by ftfont backend
* lisp/language/hebrew.el (hebrew-shape-gstring): If the font backend
supports rendering of combining characters, call
font-shape-gstring.
* src/font.c (Ffont_get): Handle `combining-capability' property.
(syms_of_font): New symbol ":combining-capability'.
* src/font.h (struct font_driver): New member combining_capability.
* src/ftfont.c: Include "category.h".
(ftfont_driver): Initialize combining_capability to
ftfont_combining_capability.
(ftfont_shape_by_flt): If OTF is null, try to find a suitable
FLT in advance.
(ftfont_combining_capability): New function.
Anders Lindgren [Tue, 26 Jan 2016 12:51:59 +0000 (13:51 +0100)]
Fixed NextStep fullscreen issue (bug#22468)
When in fullscreen mode, `[screen visibleFrame]' sometimes
includes, sometimes excludes the menu bar. This could cause
a frame to be placed too low when in fullscreen mode.
* src/nsterm.m (ns_menu_bar_should_be_hidden): Trace.
(constrain_frame_rect): New parameter, isFullscreen, when true don't
query the height of the menu bar.
(ns_constrain_all_frames): Pass `false' (isFullscreen) to
`constrain_frame_rect'.
([EmacsView initFrameFromEmacs:]): Trace.
([EmacsView isFullscreen]): Trace.
([EmacsWindow constrainFrameRect:toScreen:]): Pass fullscreen
state to `constrain_frame_rect'.
Artur Malabarba [Mon, 25 Jan 2016 22:42:50 +0000 (22:42 +0000)]
* lisp/files.el: Use a fixed file name for the second dir-locals file
(dir-locals-file): Revert to its original fixed value.
(dir-locals-file-2): New const.
(dir-locals--all-files): Don't use `file-name-all-completions'.
Instead, just check for the 2 dir-locals files and return a list
of the ones that exit (if any).
* etc/NEWS: Document the change.
* doc/emacs/custom.texi (Directory Variables): Document the change.
* doc/lispref/variables.texi (Directory Local Variables): Update
accordingly.
Returning a cache remains unchanged, but the case of returning a
file (or pattern) is now changed to return the contaning
directory.
(dir-locals-read-from-file): Rename to `dir-locals-read-from-dir'
and make obsolete.
(dir-locals-read-from-dir): Simplify accordingly.
(hack-dir-local-variables): Simplify accordingly and rename a
variable.