ChangeLog

2014-07-05  Eli Zaretskii  <eliz@gnu.org>

	* xdisp.c (pos_visible_p): If CHARPOS is at beginning of window,
	and there is a display property at that position, don't call
	move_it_to to move to a position before window start.  (Bug#17942)
	Fix condition for finding CHARPOS by the first call to move_it_to.
	(Bug#17944)

2014-07-05  Stefan Monnier  <monnier@iro.umontreal.ca>

	* syntax.c (find_defun_start): Try the cache even
	if !open_paren_in_column_0_is_defun_start.
	(back_comment): If find_defun_start was pessimistic, use the
	scan_sexps_forward result to improve the cache (bug#16526).

2014-07-04  Daniel Colascione  <dancol@dancol.org>

	Backport from trunk.
	* xfns.c (create_frame_xic): Pass XNStatusAttributes to XCreateIC
	only if xic_style calls for it.  This change allows Emacs to work
	with ibus.  Also, don't leak resources if create_frame_xic fails,
	and stop caching xic_style across different displays (Bug#17928).
	(supported_xim_styles): Make const.
	(best_xim_style): Remove first parameter: it's always just
	supported_xim_styles.  Change to look at supported_xim_styles directly.

2014-07-04  Eli Zaretskii  <eliz@gnu.org>

	* xdisp.c (redisplay_window): If redisplay of a window ends up
	with point in a partially visible line at end of the window, make
	sure the amended position of point actually has smaller Y
	coordinate; if not, give up and scroll the display.  (Bug#17905)

	* window.c (window_scroll_pixel_based): When point ends up at the
	last fully visible line, don't let move_it_to stop at the left
	edge of the line and dupe us into thinking point is inside the
	scroll margin.

	* w32.c (network_interface_info): Make sure the argument is a
	Lisp string.

2014-07-01  Eli Zaretskii  <eliz@gnu.org>

	* dispnew.c (prepare_desired_row): Accept 2 additional arguments:
	the window whose glyph row is being prepared and a flag whether it
	is for mode/header line.  Make sure the glyph row's marginal areas
	are in sync with what the window wants.
	(Bug#17892)

	* xdisp.c (display_line, display_mode_line): Call
	prepare_desired_row with additional arguments, as appropriate.

	* dispextern.h (prepare_desired_row): Adjust prototype.

2014-07-01  Dmitry Antipov  <dmantipov@yandex.ru>

	* xfaces.c (init_frame_faces): Always realize basic faces (Bug#17889).
	* menu.c (Fx_popup_dialog): Set Vmenu_updating_frame to avoid crash
	caused by xw_popup_dialog in daemon mode (Bug#17891).

2014-06-30  Eli Zaretskii  <eliz@gnu.org>

	* frame.c (do_switch_frame): When switching to another TTY frame,
	make sure FrameCols and FrameRows are in sync with the new frame's
	data.  (Bug#17875)

2014-06-28  Andreas Schwab  <schwab@linux-m68k.org>

	* coding.c (encode_coding_utf_8): Correctly count produced_chars
	also in unibyte case.  (Bug#17865)

2014-06-25  Glenn Morris  <rgm@gnu.org>

	* puresize.h (BASE_PURESIZE): Increase a bit.  (Bug#17846)

2014-06-21  Eli Zaretskii  <eliz@gnu.org>

	* indent.c (Fvertical_motion): Doc fix.
	Move to the goal column, if any, with a single call to
	move_it_in_display_line, not in two calls.  Doing this with two
	calls causes move_it_in_display_line apply the line-prefix
	handling twice instead of just once.  (Bug#17823)

2014-06-21  Paul Eggert  <eggert@cs.ucla.edu>

	Port to OS X ACLs (Bug#17810).
	* fileio.c (Ffile_acl): Port to OS X, where acl_get_file (...,
	ACL_TYPE_ACCESS) doesn't work.

2014-06-19  Stefan Monnier  <monnier@iro.umontreal.ca>

	* keyboard.c (read_key_sequence): Don't invoke Vprefix_help_command
	before checking key-translation-map (bug#17659).

2014-06-19  Dmitry Antipov  <dmantipov@yandex.ru>

	* font.c (font_make_object): Avoid dangling pointer which may
	crash GC (Bug#17771).

2014-06-18  Eli Zaretskii  <eliz@gnu.org>

	* image.c [5 < GIFLIB_MAJOR + (1 <= GIFLIB_MINOR)]: Declare the
	prototype of DGifCloseFile as appropriate for older and newer
	versions of giflib.
	(gif_close): New function, encapsulates the differences in the
	calling sequence of DGifCloseFile before v5.1.0 and after it.
	(gif_load): Call gif_close instead of DGifCloseFile.  Divulge the
	error string where appropriate.  (Bug#17790)

2014-06-16  Eli Zaretskii  <eliz@gnu.org>

	* xdisp.c (Fmove_point_visually): Instead of testing for keyboard
	macro execution, make sure point didn't move since last complete
	redisplay, as the condition for using the glyph matrix
	information.  (Bug#17777)

2014-06-14  Eli Zaretskii  <eliz@gnu.org>

	* xdisp.c (Fmove_point_visually): Don't use the glyph matrix
	information if we are in the middle of executing a keyboard macro,
	since redisplay doesn't update the screen until the macro is
	finished.  (Bug#17777)

2014-06-13  Eli Zaretskii  <eliz@gnu.org>

	* alloc.c (cleanup_vector): Don't dereference a font driver
	pointer if it is NULL.  (Bug#17771)

2014-06-11  Eli Zaretskii  <eliz@gnu.org>

	* xdisp.c (set_cursor_from_row): Fix an off-by-one error when
	matching overlay strings with 'cursor' property against buffer
	positions traversed in the glyph row.  (Bug#17744)

2014-06-08  Glenn Morris  <rgm@gnu.org>

	* fileio.c (write-region-inhibit-fsync): Doc tweak.

	* data.c (Flss, Fgtr, Fleq, Fgeq): Doc tweaks.

2014-06-04  Eli Zaretskii  <eliz@gnu.org>

	* sysdep.c (reset_sys_modes): Use cursorX, not curX, as the latter
	contains garbage on WINDOWSNT (which could potentially infloop at
	exit).

	Minimize cursor motion during TTY menu updates.
	* term.c (tty_menu_display): Don't position cursor here.  Instead,
	pass the cursor coordinates to update_frame_with_menu.
	(tty_menu_activate): Send the hide cursor command only once in an
	iteration through the outer 'while' loop.

	* dispnew.c (update_frame_1): Accept an additional argument
	SET_CURSOR_P, and position the cursor at the end of the frame
	update only if that argument is non-zero.  All callers changed to
	provide the additional argument as non-zero, except for
	update_frame_with_menu.
	(update_frame_with_menu): Accept 2 additional arguments ROW and
	COL; if they are non-negative, instruct update_frame_1 not to
	position the cursor, and instead position it according to ROW and
	COL.

	* dispextern.h (update_frame_with_menu): Update prototype.

2014-06-02  Stefan Monnier  <monnier@iro.umontreal.ca>

	* callproc.c (call_process): Don't check read-only if we don't insert
	anything (bug#17666).

2014-06-02  Eli Zaretskii  <eliz@gnu.org>

	* dispnew.c (update_frame_with_menu): Set display_completed.

2014-06-01  Paul Eggert  <eggert@cs.ucla.edu>

	* frame.c (x_set_frame_parameters): Don't read uninitialized storage.

2014-06-01  Jan Djärv  <jan.h.d@swipnet.se>

	* xterm.c (xg_scroll_callback): Remove position, for jump set portion
	to min(value, whole).

2014-06-01  Paul Eggert  <eggert@cs.ucla.edu>

	Bring back the changes to GDB-visible symbols, but only on AIX.
	And only if it's not pre-4.2 GCC.
	* lisp.h (ENUMABLE, DEFINE_GDB_SYMBOL_ENUM): New macros.
	(ARRAY_MARK_FLAG, PSEUDOVECTOR_FLAG, VALMASK): Use them.
	(ARRAY_MARK_FLAG_val, PSEUDOVECTOR_FLAG_val, VALMASK_val):
	New macros.

2014-05-31  Eli Zaretskii  <eliz@gnu.org>

	* fileio.c (Finsert_file_contents): Call prepare_to_modify_buffer
	with PT, not GPT.  (Bug#16433)

	Revert last changes to GDB-visible symbols.
	* lisp.h (ENUMABLE, DEFINE_GDB_SYMBOL_ENUM): Delete macros.
	(ARRAY_MARK_FLAG, PSEUDOVECTOR_FLAG, VALMASK): Don't use them.
	(ARRAY_MARK_FLAG_val, PSEUDOVECTOR_FLAG_val, VALMASK_val):
	Delete macros.

2014-05-31  Glenn Morris  <rgm@gnu.org>

	* cmds.c (Fself_insert_command): Allow zero repeat count.  (Bug#17649)

2014-05-30  Paul Eggert  <eggert@cs.ucla.edu>

	Fix port to 32-bit AIX with xlc (Bug#17598).
	* alloc.c (gdb_make_enums_visible): Remove FLOAT_TO_STRING_BUFSIZE.
	* conf_post.h (FLEXIBLE_ARRAY_MEMBER) [__IBMC__]: Don't define to empty.
	* lisp.h (FLOAT_TO_STRING_BUFSIZE): Make it a macro, instead of an enum,
	to work around a compiler bug in IBM xlc 12.1.

2014-05-29  Eli Zaretskii  <eliz@gnu.org>

	* xterm.c (x_update_window_end): Don't invalidate the entire
	mouse-highlight info, just signal frame_up_to_date_hook that mouse
	highlight needs to be redisplayed.  (Bug#17588)

2014-05-29  Paul Eggert  <eggert@cs.ucla.edu>

	Port the GDB-visible symbols to AIX.
	Without them, GDB doesn't work to debug Emacs, since the AIX linker
	optimizes away the relevant external symbols.  Use enums instead;
	this suffices for the AIX port, which is 32-bit-only anyway.
	* lisp.h (ENUMABLE, DEFINE_GDB_SYMBOL_ENUM): New macros.
	(ARRAY_MARK_FLAG, PSEUDOVECTOR_FLAG, VALMASK): Use them.
	(ARRAY_MARK_FLAG_val, PSEUDOVECTOR_FLAG_val, VALMASK_val):
	New macros.

2014-05-26  Paul Eggert  <eggert@cs.ucla.edu>

	Include sources used to create macuvs.h.
	* Makefile.in ($(srcdir)/macuvs.h): New rule.
	* macuvs.h: Use automatically-generated header.

2014-05-25  Eli Zaretskii  <eliz@gnu.org>

	* xdisp.c (move_it_in_display_line_to): Don't record wrap position
	if we are iterating over an object that generates glyphs for
	marginal areas.  (Bug#17585)

2014-05-24  Paul Eggert  <eggert@cs.ucla.edu>

	* xdisp.c (safe__call1, safe__eval): Now static.

2014-05-24  Eli Zaretskii  <eliz@gnu.org>

	* xdisp.c (safe__call): Accept va_list argument instead of '...'.
	(safe_call, safe__call1): Construct a va_list argument for safe_call.
	(safe_call1): Call safe_call instead of safe__call directly.

2014-05-24  Ken Brown  <kbrown@cornell.edu>

	* w32term.c (x_delete_display) [CYGWIN]: Don't free
	dpyinfo->w32_id_name, to make sure it doesn't get freed more than
	once.  (Bug#17510)

2014-05-24  Stefan Monnier  <monnier@iro.umontreal.ca>

	* xdisp.c: Bind inhibit-quit during pre-redisplay-function.
	(safe__call, safe__call1, safe__eval): New functions.
	(safe_call): Use it.
	(prepare_menu_bars): Use it for pre-redisplay-function (bug#17577).
	(display_mode_element): Same for `:eval'.

2014-05-22  Paul Eggert  <eggert@cs.ucla.edu>

	Fix port to 32-bit AIX (Bug#17540).
	* unexaix.c (copy_text_and_data): Don't add text_scnptr to ptr
	twice.  _text already includes this offset.
	(unrelocate_symbols): Don't cast 64-bit integer to pointer.

2014-05-21  Eli Zaretskii  <eliz@gnu.org>

	* xdisp.c (move_it_in_display_line_to): Avoid infinite recursion:
	when closest_pos is identical to to_charpos, don't recurse, since
	we already tried that, and failed.  (Bug#17539)

2014-05-20  Eli Zaretskii  <eliz@gnu.org>

	* w32fns.c (unwind_create_frame) [GLYPH_DEBUG]: If we are
	unwinding when frame's faces were not initialized yet, increment
	the frame's image-cache reference count before calling
	x_free_frame_resources.  Don't dereference
	dpyinfo->terminal->image_cache if it is NULL.  (Bug#17524)

2014-05-11  Glenn Morris  <rgm@gnu.org>

	* fileio.c (Ffile_executable_p): Doc tweak.

2014-05-10  Jan Djärv  <jan.h.d@swipnet.se>

	* xsettings.c (init_gsettings): Use g_settings_schema_source_lookup
	instead of deprecated g_settings_list_schemas if possible (Bug#17434).

2014-05-07  Paul Eggert  <eggert@cs.ucla.edu>

	* minibuf.c (read_minibuf): Avoid C99ism in previous patch (Bug#17430).

2014-05-07  Jarek Czekalski  <jarekczek@poczta.onet.pl>

	Fix initialization of minibuffer history variable (Bug#17430).
	* minibuf.c (read_minibuf): Initialize histval to Qnil if unbound.
	Move the initialization up to prevent any "value void" message.

2014-05-06  Samuel Bronson  <naesten@gmail.com>

	* keyboard.c (Frecursive_edit): Ensure inc&dec of command_loop_level
	are matched (bug#17413).

2014-05-06  Jarek Czekalski  <jarekczek@poczta.onet.pl>

	Stop tooltips pulling Emacs window to front (Bug#17408).
	* w32fns.c (Fx_show_tip): Add SWP_NOOWNERZORDER flag to
	SetWindowPos invocations.

2014-05-05  Jan Djärv  <jan.h.d@swipnet.se>

	* nsselect.m (Fx_selection_exists_p): Just return Qnil if window system
	not initialized (Bug#17398).

2014-05-04  Paul Eggert  <eggert@cs.ucla.edu>

	Revert recent libpng changes (Bug#17339).
	* Makefile.in (PNG_CFLAGS): Remove; all uses removed.
	* image.c [HAVE_LIBPNG_PNG_H]: Include <libpng/png.h>, not <png.h>.

2014-05-03  Paul Eggert  <eggert@cs.ucla.edu>

	Handle systems without WCONTINUED consistently.  (Bug#15110, 17339)
	* process.c (handle_child_signal): Remove WCONTINUED ifdef,
	because WCONTINUED is always defined now.
	* syswait.h (WCONTINUED): Move here from ../lib-src/emacsclient.c.

2014-05-03  Eli Zaretskii  <eliz@gnu.org>

	* buffer.c (overlay_strings): Fix the wording of the commentary.

2014-05-02  Paul Eggert  <eggert@cs.ucla.edu>

	Consult libpng-config more consistently (Bug#17339).
	* Makefile.in (PNG_CFLAGS): New var.
	(ALL_CFLAGS): Use it.
	* image.c [HAVE_PNG]: Don't worry about <libpng/png.h>, as
	CFLAGS now handles this.

2014-05-01  Glenn Morris  <rgm@gnu.org>

	* floatfns.c (Fisnan):
	* profiler.c (Fprofiler_cpu_running_p): Doc fix (replace `iff').

2014-05-01  Eli Zaretskii  <eliz@gnu.org>

	* term.c (tty_menu_activate): A better initialization for cursor
	coordinates.

2014-05-01  Stefan Monnier  <monnier@iro.umontreal.ca>

	* intervals.c: Tighten assertions.
	(create_root_interval): Make sure the interval is not empty.
	(intervals_equal): Use booleans.
	(rotate_right, rotate_left): Check LENGTHs rather than TOTAL_LENGTH.
	(balance_an_interval): Sanity check LENGTHs and TOTAL_LENGTHs.
	(balance_possible_root_interval): Simplify and use booleans.
	(split_interval_right, split_interval_left): Check LENGTH, and remove
	now redundant assertion.
	(adjust_intervals_for_insertion): Remove now redundant assertions.
	(delete_node, interval_deletion_adjustment)
	(adjust_intervals_for_deletion, merge_interval_right)
	(merge_interval_left): Check LENGTH rather than TOTAL_LENGTH.
	(reproduce_interval): Make sure the interval is not empty.

2014-04-30  Paul Eggert  <eggert@cs.ucla.edu>

	* term.c (tty_menu_activate): Don't assume row and col are initialized.
	GCC 4.9.0 warned about this, and I couldn't easily prove to my own
	satisfaction that they would always be initialized.

2014-04-29  Eli Zaretskii  <eliz@gnu.org>

	* term.c (tty_menu_display): Move the cursor to the active menu
	item.
	(tty_menu_activate): Return the cursor to the active menu item
	after displaying the menu and after displaying help-echo.  See
	http://lists.gnu.org/archive/html/emacs-devel/2014-04/msg00402.html
	for the details of why this is needed by screen readers and
	Braille displays.

2014-04-29  Glenn Morris  <rgm@gnu.org>

	* process.c (handle_child_signal):
	Handle systems without WCONTINUED.  (Bug#15110, 17339)

2014-04-23  Eli Zaretskii  <eliz@gnu.org>

	* search.c (Fnewline_cache_check): Don't try to count newlines
	outside the buffer's restriction, as find_newline doesn't support that.

2014-04-22  Paul Eggert  <eggert@cs.ucla.edu>

	Port to GCC 4.9.0 with --enable-gcc-warnings.
	* image.c (struct my_jpeg_error_mgr) [lint]: Remove member fp.
	All uses removed.
	(jpeg_load_body) [lint]: Add a 'volatile' to pacify a buggy GCC in
	a way that also works with GCC 4.9.0.

	* search.c (Fnewline_cache_check): Remove unused locals.

2014-04-22  Eli Zaretskii  <eliz@gnu.org>

	* search.c (find_newline1): New subroutine.
	(Fnewline_cache_check): New function.
	(syms_of_search): Defsubr it.

2014-04-21  Jarek Czekalski  <jarekczek@poczta.onet.pl>

	Fix freezing with scroll bars of GTK3 Toolkit (bug#15801).
	* keyboard.c (unblock_input): Add comment.
	* xgselect.c (xg_select): Prevent Glib main loop recursion.

2014-04-19  Stefan Monnier  <monnier@iro.umontreal.ca>

	* intervals.c (rotate_right, rotate_left): Fix up length computation.
	Also change identifiers to match the comments, and add more assertions
	(bug#16234).

2014-04-18  Eli Zaretskii  <eliz@gnu.org>

	* xdisp.c (insert_left_trunc_glyphs): Ensure the left truncation
	glyph is written to TEXT_AREA of the temporary glyph_row.
	(Bug#17288)

2014-04-18  Paul Eggert  <eggert@cs.ucla.edu>

	* emacs.c (close_output_streams): Don't clear and restore errno.

2014-04-17  Jan Djärv  <jan.h.d@swipnet.se>

	* xterm.c (x_make_frame_visible): Prevent endless loop when frame
	never becomes visible, i.e. using XMonad  (Bug#17237).

2014-04-17  Eli Zaretskii  <eliz@gnu.org>

	* xdisp.c (Fline_pixel_height): Don't assume that the current
	buffer and the selected window's buffer are one and the same.
	(Bug#17281)

	* insdel.c (invalidate_buffer_caches): Invalidate the bidi
	paragraph-start cache before the newline cache.  (Bug#17269)

2014-04-16  Stefan Monnier  <monnier@iro.umontreal.ca>

	* window.c (save_window_save): Lookup window_point_insertion_type in
	the right buffer (bug#15457).
	(Qwindow_point_insertion_type): New var.
	(syms_of_window): Initialize it.

2014-04-16  Eli Zaretskii  <eliz@gnu.org>

	Fix the MSDOS build.
	* unexcoff.c [MSDOS]: Include libc/atexit.h.
	(copy_text_and_data): Zero out the atexit chain pointer before
	dumping Emacs.

	* termhooks.h (encode_terminal_code): Update prototype.

	* term.c (encode_terminal_code) [DOS_NT]: Make it externally
	visible for all DOS_NT ports, not just WINDOWSNT.
	(syms_of_term) [!MSDOS]: Don't define 'tty-menu-*' symbols on MSDOS.

	* sysdep.c (emacs_sigaction_init, init_signals): Don't use SIGCHLD
	unless it is defined.
	(emacs_pipe) [MSDOS]: Redirect to 'pipe'.

	* process.c (close_on_exec, accept4, process_socket): Move into
	the "ifdef subprocesses" part.
	(catch_child_signal): Condition by "ifdef subprocesses".
	(syms_of_process) <Qinternal_default_process_sentinel>
	<Qinternal_default_process_filter>: Condition by "ifdef subprocesses".

	* msdos.h: Add prototypes for new functions.
	(EINPROGRESS): Define.
	(O_CLOEXEC): Define to zero.

	* msdos.c (check_window_system): Remove unnecessary an
	incompatible duplicate function.
	(sys_opendir, readlinkat, faccessat, fstatat, unsetenv):
	New functions in support of new functionality.

	* menu.c (single_menu_item): Add visual indication  of submenu
	also for menus on MSDOS frames.
	(Fx_popup_menu) [!MSDOS]: Do not call tty_menu_show on MSDOS.

	* lisp.h (CHECK_PROCESS) [!subprocesses]: Do not define
	when async subprocesses aren't supported.

	* font.h (FONT_WIDTH) [MSDOS]: MSDOS-specific definition.

	* emacs.c (close_output_streams): Zero out errno before calling
	close_stream.

	* dired.c [MSDOS]: Include msdos.h.

	* conf_post.h (opendir) [MSDOS]: Redirect to sys_opendir.
	(DATA_START) [MSDOS]: Define.
	(SYSTEM_PURESIZE_EXTRA) [MSDOS]: Enlarge by 25K.

	* callproc.c (block_child_signal, unblock_child_signal) [MSDOS]:
	Ifdef away for MSDOS.
	(record_kill_process) [MSDOS]: Ifdef away the entire body for MSDOS.
	(call_process_cleanup) [MSDOS]: Ifdef away portions not relevant
	for MSDOS.
	(call_process) [MSDOS]: Fix call sequence of dostounix_filename.
	Use temporary file template that is compatible with mkostemp.
	Move vfork-related portions under #ifndef MSDOS.
	(syms_of_callproc): Unify templates of MSDOS and WINDOWSNT.

2014-04-15  Stefan Monnier  <monnier@iro.umontreal.ca>

	* buffer.c (Foverlays_at): Add argument `sorted'.

2014-04-14  Eli Zaretskii  <eliz@gnu.org>

	* insdel.c (invalidate_buffer_caches): When deleting or replacing
	text, invalidate the bidi_paragraph_cache upto and including the
	preceding newline.

2014-04-13  Paul Eggert  <eggert@cs.ucla.edu>

	Port to IRIX 6.5 (Bug#9684).
	* alloc.c (TAGGABLE_NULL): New constant,
	for porting to hosts with nontrivial DATA_SEG_BITS settings.
	(next_vector, set_next_vector): Use it.
	* conf_post.h (INET6) [IRIX6_5]: Define.
	(HAVE_GETADDRINFO) [IRIX6_5]: Undef.
	* data.c (BITS_PER_ULL): Don't assume ULLONG_MAX is defined.
	* lisp.h (lisp_h_XPNTR): Don't OR in bits that aren't masked out,
	for consistency with how TAGGABLE_NULL is computed.

2014-04-13  Eli Zaretskii  <eliz@gnu.org>

	* keyboard.c (Fopen_dribble_file): Encode the dribble file-name
	before passing it to system APIs.

	* puresize.h (BASE_PURESIZE): Bump by 1K.  (Bug#17255)

2014-04-13  Stefan Monnier  <monnier@iro.umontreal.ca>

	* bytecode.c (exec_byte_code): Rework the volatiles.  Most importantly,
	make sure stack.byte_string_start is not de-adjusted by pushhandler.

2014-04-12  Paul Eggert  <eggert@cs.ucla.edu>

	* keyboard.c (Fopen_dribble_file): Avoid some races.  (Bug#17187)

2014-04-12  Eli Zaretskii  <eliz@gnu.org>

	* xdisp.c (move_it_by_lines): If a large portion of buffer text is
	covered by a display string that ends in a newline, and that cases
	going back by DVPOS lines to hit the search limit, lift the limit
	and go back until DVPOS is reached.  (Bug#17244)

	* indent.c (Fvertical_motion): Handle correctly the case when the
	display string is preceded by an empty line.

2014-04-11  Eli Zaretskii  <eliz@gnu.org>

	* w32.c (sys_umask) <WRITE_USER>: Remove redundant constant, and
	use S_IWRITE instead.

2014-04-11  Glenn Morris  <rgm@gnu.org>

	* keyboard.c (Fopen_dribble_file): Make file private.  (Bug#17187)

2014-04-09  Ken Brown  <kbrown@cornell.edu>

	* Makefile.in (EMACS_MANIFEST): Revert last change.

2014-04-07  Stefan Monnier  <monnier@iro.umontreal.ca>

	* lisp.h (struct Lisp_Symbol): New bitfield `pinned'.

	* alloc.c: Keep track of symbols referenced from pure space (bug#17168).
	(symbol_block_pinned): New var.
	(Fmake_symbol): Initialize `pinned'.
	(purecopy): New function, extracted from Fpurecopy.  Mark symbols as
	pinned and signal an error for un-purifiable objects.
	(pure_cons): Use it.
	(Fpurecopy): Use it, except for objects that can't be purified.
	(mark_pinned_symbols): New function.
	(Fgarbage_collect): Use it.
	(gc_sweep): Remove hack made unnecessary.

2014-04-05  Glenn Morris  <rgm@gnu.org>

	* keyboard.c (Fopen_dribble_file): Doc tweak.

2014-04-04  Jan Djärv  <jan.h.d@swipnet.se>

	Backport from trunk.
	* nsterm.m (updateFrameSize:): If waiting for the tool bar and tool
	bar is zero height, just return (Bug#16976).
	(initFrameFromEmacs:): Initialize wait_for_tool_bar.
	* nsterm.h (EmacsView): Add wait_for_tool_bar.
	* nsmenu.m (update_frame_tool_bar): Return early if view or toolbar
	is nil.  If waiting for toolbar to complete, force a redraw.
	(free_frame_tool_bar): Set wait_for_tool_bar = NO (Bug#16976).

2014-04-03  Ken Brown  <kbrown@cornell.edu>

	* Makefile.in (EMACS_MANIFEST): Update comment.  (Bug#17176)

2014-04-02  Martin Rudalics  <rudalics@gmx.at>

	* xterm.c (x_new_font): Don't calculate non-toolkit scrollbar
	width from font width (Bug#17163).

2014-03-30  Martin Rudalics  <rudalics@gmx.at>

	* frame.c (x_set_frame_parameters): Calculate default values of
	new frame sizes only after all other frame parameters have been
	processed (Bug#17142).

2014-03-28  Ken Brown  <kbrown@cornell.edu>

	* conf_post.h (SYSTEM_PURESIZE_EXTRA) [CYGWIN]: Set to 10000.
	(Bug#17112)

2014-03-28  YAMAMOTO Mitsuharu  <mituharu@math.s.chiba-u.ac.jp>

	* xterm.c (x_draw_image_glyph_string): Adjust image background
	width accordingly when its x position is adjusted.  (Bug#17115)

2014-03-27  YAMAMOTO Mitsuharu  <mituharu@math.s.chiba-u.ac.jp>

	* w32term.c (x_draw_image_glyph_string): Fix computation of height
	and width of image background when it is displayed with a 'box'
	face.  (Bug#17115)

2014-03-26  Paul Eggert  <eggert@penguin.cs.ucla.edu>

	More backward-compatible fix to char-equal core dump (Bug#17011).
	* editfns.c (Fchar_equal): In unibyte buffers, assume values in
	range 128-255 are raw bytes.  Suggested by Eli Zaretskii.

	Fix core dump in char-equal (Bug#17011).
	* editfns.c (Fchar_equal): Do not use MAKE_CHAR_MULTIBYTE in
	unibyte buffers, as we can't tell whether the characters are
	actually unibyte.

	* insdel.c (adjust_markers_for_delete): Remove unused local.

2014-03-24  Barry O'Reilly  <gundaetiapo@gmail.com>

	Have (MARKER . ADJUSTMENT) undo records always be immediately
	after their corresponding (TEXT . POS) record in undo list.
	(Bug#16818)
	* lisp.h (record-delete): New arg record_markers.
	(record_marker_adjustment): No longer needed outside undo.c.
	* insdel.c (adjust_markers_for_delete): Move calculation of marker
	adjustments to undo.c's record_marker_adjustments.  Note that
	fileio.c's decide_coding_unwind is another caller to
	adjust_markers_for_delete.  Because it has undo list bound to t,
	it does not rely on adjust_markers_for_delete to record marker
	adjustments.
	(del_range_2): Swap call to record_delete and
	adjust_markers_for_delete so as undo marker adjustments are
	recorded before current deletion's adjustments, as before.
	(adjust_after_replace):
	(replace_range): Pass value for new record_markers arg to
	delete_record.
	* undo.c (record_marker_adjustment): Rename to
	record_marker_adjustments and made static.
	(record_delete): Check record_markers arg and call
	record_marker_adjustments.
	(record_change): Pass value for new record_markers arg to
	delete_record.
	(record_point): at_boundary calculation no longer needs to account
	for marker adjustments.

2014-03-24  Martin Rudalics  <rudalics@gmx.at>

	* w32term.c (x_set_window_size): Refine fix from 2014-03-14
	(Bug#17077).

2014-03-23  Glenn Morris  <rgm@gnu.org>

	* fileio.c (Ffile_symlink_p): Doc fix. (Bug#17073)

2014-03-23  Stefan Monnier  <monnier@iro.umontreal.ca>

	* buffer.c (struct sortvec): Add field `spriority'.
	(compare_overlays): Use it.
	(sort_overlays): Set it.

2014-03-23  Eli Zaretskii  <eliz@gnu.org>

	* xdisp.c (redisplay_window): If all previous attempts to find the
	cursor row failed, try a few alternatives before falling back to
	the top-most row of the window.  Use row_containing_pos.  (Bug#17047)

2014-03-22  Daniel Colascione  <dancol@dancol.org>

	* process.c (conv_sockaddr_to_lisp): When extracting the string
	names of AF_LOCAL sockets, stop before reading uninitialized
	memory.

2014-03-23  Daniel Colascione  <dancol@dancol.org>

	* process.c (DATAGRAM_CONN_P): Don't underflow datagram_address
	array.  (ASAN caught.)

2014-03-21  YAMAMOTO Mitsuharu  <mituharu@math.s.chiba-u.ac.jp>

	Fix regression introduced by patch for Bug#10500.
	* xterm.c (x_draw_image_relief): Respect Vtool_bar_button_margin.
	* w32term.c (x_draw_image_relief): Likewise.

2014-03-21  Martin Rudalics  <rudalics@gmx.at>

	* w32fns.c (w32_wnd_proc): For WM_WINDOWPOSCHANGING don't
	constrain frame size in SW_SHOWMAXIMIZED case so we can truly
	maximize a frame for odd default fonts.

2014-03-20  Paul Eggert  <eggert@cs.ucla.edu>

	* doc.c (store_function_docstring): Fix pointer signedness mismatch.

2014-03-20  Stefan Monnier  <monnier@iro.umontreal.ca>

	* doc.c (store_function_docstring): Warn when we don't know where to
	put a docstring.
	(Fsubstitute_command_keys): Don't advertise the fact that
	text-properties are dropped, since we think it's a bug that we'll fix
	in 24.5.

	* frame.h (SET_FRAME_VISIBLE): Keep frame_garbaged up to date.
	* xterm.c (handle_one_xevent) <MapNotify>: Don't garbage the frame.
	* frame.c (frame_garbaged): Make "docstring" more precise.

2014-03-20  Glenn Morris  <rgm@gnu.org>

	* charset.c (init_charset): When we cannot find the charsets directory,
	mention if EMACSDATA is set.

2014-03-19  Paul Eggert  <eggert@cs.ucla.edu>

	* fns.c (Frandom): Fix rare bug where the result isn't random.

	Fix porting inconsistency about rounding to even.
	* floatfns.c (emacs_rint) [!HAVE_RINT]: Round to even.
	This way, the unusual !HAVE_RINT case acts like the usual
	HAVE_RINT case, and we can fix the documentation accordingly.

2014-03-19  Eli Zaretskii  <eliz@gnu.org>

	* w32fns.c (reset_modifiers): Zero out keystate[] before using it.
	(w32_wnd_proc): Initialize the dwHoverTime member of
	TRACKMOUSEEVENT structure.

2014-03-17  Teodor Zlatanov  <tzz@lifelogs.com>

	* gnutls.c (Fgnutls_boot): Fix case of :verify-error = t.

2014-03-16  Eli Zaretskii  <eliz@gnu.org>

	* search.c (find_newline): Speed up the function when using the
	newline cache, by halving the number of calls to
	region_cache_forward and region_cache_backward.  (Bug#16830)

2014-03-15  Juanma Barranquero  <lekktu@gmail.com>

	* buffer.c (Fset_buffer): Document return value (bug#17015).

2014-03-14  Martin Rudalics  <rudalics@gmx.at>

	* w32term.c (x_set_window_size): When frame-resize-pixelwise is
	nil, always resize character wise to avoid potential loss of the
	mode line (Bug#16923 related).

2014-03-12  Martin Rudalics  <rudalics@gmx.at>

	* frame.c (x_set_frame_parameters): Always calculate new sizes
	pixelwise to avoid potential loss when rounding.

2014-03-11  Dmitry Antipov  <dmantipov@yandex.ru>

	* xfns.c (x_set_mouse_color): Recolor vertical_drag_cursor.
	* xterm.c (x_free_frame_resources): Free all allocated cursors.

2014-03-10  Eli Zaretskii  <eliz@gnu.org>

	* w32.c (fstatat): Don't add an extra slash if the argument ends
	with a slash: this fails the subsequent call to stat_worker on
	Windows 9X.  Reported by oslsachem <oslsachem@gmail.com>.

2014-03-09  Martin Rudalics  <rudalics@gmx.at>

	* xdisp.c (Fwindow_text_pixel_size): Adjust doc-string.

2014-03-08  Jan Djärv  <jan.h.d@swipnet.se>

	* nsterm.h (MAC_OS_X_VERSION_10_9): Add.

	* nsterm.m (constrainFrameRect:toScreen:): Constrain normally
	when frame is only on one screen (Bug#14713).

2014-03-08  Eli Zaretskii  <eliz@gnu.org>

	* xdisp.c (move_it_in_display_line_to): If word-wrap is ON, and
	there's a valid wrap point in the display line, the last glyph
	cannot "just barely fit" on this row, because display_line doesn't
	let it.  Instead, proceed as if the last glyph didn't fit, so that
	we eventually back up the iterator to the wrap point.  This avoids
	delusional behavior of move_it_to, whereby it proceeds to the next
	display line, but sets current_x to zero for all the glyphs that
	without word-wrap would fit on the previous display line.
	One result was that visual-order cursor movement behaved erratically
	under word-wrap.
	(Fmove_point_visually): Add code to find the x coordinate of the
	last character before wrap point, under word-wrap on a TTY.

2014-03-07  Eli Zaretskii  <eliz@gnu.org>

	* xdisp.c (Fmove_point_visually): When under word-wrap, accept
	also return value of MOVE_POS_MATCH_OR_ZV from
	move_it_in_display_line_to, when moving from beginning of line to
	point's position.  (Bug#16961)

2014-03-07  Martin Rudalics  <rudalics@gmx.at>

	* buffer.c (Vbuffer_list_update_hook): Doc-string fix.
	* window.c (Fselect_window): Explain NORECORD and
	`buffer-list-update-hook' in doc-string.

2014-03-06  Martin Rudalics  <rudalics@gmx.at>

	* window.c (Fother_window_for_scrolling): Check that
	Vother_window_scroll_buffer is a buffer.

2014-03-06  Dmitry Antipov  <dmantipov@yandex.ru>

	* xterm.c (xim_initialize): Always pass a copy of resource name
	to XRegisterIMInstantiateCallback and eassert whether return
	value is True.  Passing copy is important because Xlib doesn't
	make its own copy and resource name argument usually points to
	SSDATA (Vx_resource_name), which may be changed from Lisp.
	(xim_close_display): For XUnregisterIMInstantiateCallback,
	always eassert return value and pass exactly the same values
	as were used for XRegisterIMInstantiateCallback.
	Otherwise XUnregisterIMInstantiateCallback will always fail.  See Xlib
	sources to check why if you are interested.

2014-03-05  Martin Rudalics  <rudalics@gmx.at>

	* dispnew.c (change_frame_size_1): Add new_lines instead of
	new_height, the latter may be still zero if passed as such.
	* window.c (Fwindow_pixel_height): Mention bottom divider in
	doc-string.

2014-03-05  Paul Eggert  <eggert@cs.ucla.edu>

	Fix "resource temporarily unavailable" with xgselect
	(Bug#16925).
	* xgselect.c: Include <stdbool.h>.
	(xg_select) [!USE_GTK]: Don't lose track of errno.

	Fix minor --enable-gcc-warnings issues.
	* widget.c (update_various_frame_slots, EmacsFrameResize):
	Avoid unused locals.  Prefer 'if' to '#if' when either will do.

2014-03-04  Ken Brown  <kbrown@cornell.edu>

	* gmalloc.c (aligned_alloc): Clarify the code by making `adj'
	represent the actual adjustment needed for alignment.

2014-03-04  Eli Zaretskii  <eliz@gnu.org>

	* gmalloc.c (aligned_alloc): Don't allocate more memory than
	needed, and don't reallocate if the initial allocation already
	fits the bill.  Suggested by Ken Brown <kbrown@cornell.edu>.

2014-03-04  YAMAMOTO Mitsuharu  <mituharu@math.s.chiba-u.ac.jp>

	* xterm.c (x_draw_stretch_glyph_string): Reset clipping.
	(Bug#16932)

2014-03-04  Michal Nazarewicz  <mina86@mina86.com>

	* cmds.c (delete-char): Update docstring pointing out that the
	function ignores `delete-active-region' and `overwrite-mode'.

2014-03-03  Eli Zaretskii  <eliz@gnu.org>

	* font.c (Fframe_font_cache): Fix last change.  (Bug#16930)

	* gmalloc.c (aligned_alloc): Fix adjustment of size of the
	allocated buffer due to alignment.
	(freehook): If the block to be freed was allocated by
	'aligned_alloc', find its real pointer before calling 'free'.
	(Bug#16901)
	(mabort) [emacs]: Call 'emacs_abort', not 'abort', to provide a
	backtrace.

2014-03-03  Dmitry Antipov  <dmantipov@yandex.ru>

	* font.c (toplevel): Adjust comment about font cache layout.
	(font_clear_cache): Fix to match real font cache layout.
	Suggested by <namespace_collision@yahoo.com> in Bug#16069.
	(Fframe_font_cache) [FONT_DEBUG]: New function.
	(syms_of_font) [FONT_DEBUG]: Defsubr it.

	Avoid crashes when X fonts are erroneously freed on reused X
	'Display *' connection data (Bug#16069).  Note that X font
	resources still may be leaked, but currently there is no way
	to completely avoid it.
	* xterm.h (struct x_display_info): New member x_id.  Add comments.
	* xterm.c (x_display_id): New variable.
	(x_term_init): Assign identifier to each opened X connection.
	* xfont.c (struct xfont): New member x_display_id.
	(xfont_open): Initialize it with frame's display id.
	(xfont_close): Check whether font's display id matches the one
	recorded for the given display.  Adjust comment.
	* xftfont.c (struct xftfont_info):
	(xftfont_open, xftfont_close): Exactly as above with xfont stuff.

2014-03-01  Martin Rudalics  <rudalics@gmx.at>

	Consider Vother_window_scroll_buffer valid iff it's a live buffer.
	* window.c (Fother_window_for_scrolling): Don't try to scroll a
	killed Vother_window_scroll_buffer.
	(Vother_window_scroll_buffer): Fix doc-string accordingly.

2014-03-01  Eli Zaretskii  <eliz@gnu.org>

	* fileio.c (Fexpand_file_name) [WINDOWSNT]: Don't treat file names
	that start with more than 2 slashes as UNCs.  (Bug#16751)

2014-02-28  Paul Eggert  <eggert@penguin.cs.ucla.edu>

	Fix a few crashes and leaks when cloning C strings.
	* alloc.c, lisp.h (dupstring): New function.
	* gtkutil.c (xg_get_font):
	* term.c (tty_default_color_capabilities):
	* xsettings.c (store_monospaced_changed)
	(store_font_name_changed, parse_settings)
	(read_and_apply_settings, init_gsettings, init_gconf): Use it.
	This avoids some unlikely crashes due to accessing freed storage,
	and avoids some minor memory leaks in the more-typical case.

2014-02-28  Martin Rudalics  <rudalics@gmx.at>

	* xdisp.c (note_mode_line_or_margin_highlight): Don't show drag
	cursor when modeline can't be dragged (Bug#16647).

2014-02-28  Glenn Morris  <rgm@gnu.org>

	* doc.c (Fsnarf_documentation): Snarf not-yet-bound variables
	from custom-delayed-init-variables.  (Bug#11565)

2014-02-27  Martin Rudalics  <rudalics@gmx.at>

	More fixes for mouse glyph calculations (Bug#16647).
	* window.c (coordinates_in_window): In intersection of
	horizontal and vertical window dividers prefer the horizontal
	one.  Add some extra parens to last fix.
	(window_relative_x_coord): Return x-coordinate for header and
	mode line too.
	* xdisp.c (remember_mouse_glyph): In text area don't extend
	glyph into mode line to show the vertical drag cursor there
	immediately.  Subdivide mouse glyphs in right fringes to show a
	horizontal drag cursor as soon as we enter the "grabbable width"
	portion.  Handle vertical border case separately.  Do not
	subdivide window divider areas.
	(note_mouse_highlight): On bottom divider of bottommost windows
	show vertical drag cursor only when the minibuffer window can be
	resized.

2014-02-27  Eli Zaretskii  <eliz@gnu.org>

	* xdisp.c (pop_it): Restore the it->face_box_p flag which could be
	reset by the face of the object just displayed.  See also bug#76.
	(get_next_display_element): If the string came from a display
	property, examine the box face attribute at it->position, not at
	it->current.pos, since the latter was not updated yet.  (Bug#16870)
	(handle_face_prop): Improve commentary.

2014-02-27  Michael Albinus  <michael.albinus@gmx.de>

	* dbusbind.c (Fdbus__init_bus, Qdbus__init_bus, Sdbus__init_bus):
	Rename from Fdbus_init_bus_1, Qdbus_init_bus_1, Sdbus_init_bus_1.

2014-02-26  Martin Rudalics  <rudalics@gmx.at>

	Fixes around Bug#16647.
	* xdisp.c (remember_mouse_glyph): Handle ON_RIGHT_DIVIDER and
	ON_BOTTOM_DIVIDER cases.
	* window.c (coordinates_in_window): Return ON_VERTICAL_BORDER
	only if the window has no right divider.
	(Fcoordinates_in_window_p): Fix doc-string.

2014-02-25  Juanma Barranquero  <lekktu@gmail.com>

	* lread.c (Funintern): Fix doc to match advertised calling convention.

2014-02-24  Daniel Colascione  <dancol@dancol.org>

	* keyboard.c (read_char): Close race that resulted in lost events.

2014-02-22  Glenn Morris  <rgm@gnu.org>

	* frame.c (frame-alpha-lower-limit, frame-resize-pixelwise):
	* window.c (window-resize-pixelwise): Doc fixes.

	* process.c (Finternal_default_process_filter)
	(Finternal_default_process_sentinel): Doc tweaks.

2014-02-21  Glenn Morris  <rgm@gnu.org>

	* process.c (Fprocess_buffer, Faccept_process_output)
	(Finternal_default_process_filter, Finternal_default_process_sentinel):
	Doc fixes.

2014-02-21  Martin Rudalics  <rudalics@gmx.at>

	* window.c (Fwindow_scroll_bar_width): New function.

2014-02-21  Paul Eggert  <eggert@cs.ucla.edu>

	Pacify GCC when configuring with --enable-gcc-warnings.
	* xdisp.c (move_it_in_display_line_to) [lint]:
	Initialize recently-added local.

2014-02-21  Daniel Colascione  <dancol@dancol.org>

	* dbusbind.c: Rename dbus-init-bus to dbus-init-bus-1.

2014-02-20  Eli Zaretskii  <eliz@gnu.org>

	* xdisp.c (init_iterator): Don't dereference a bogus face
	pointer.  (Bug#16819)
	(try_cursor_movement): Don't use cursor position if
	set_cursor_from_row failed to compute it.  This avoids assertion
	violations in MATRIX_ROW.
	(move_it_in_display_line_to): Save the iterator state in ppos_it
	only once per call.  Reimplement the method used to return to the
	best candidate position if all the positions found in display line
	are beyond TO_CHARPOS.  This cuts down the number of calls to
	bidi_shelve_cache, which moves a lot of stuff when lines are long
	and include bidirectional text.  (Bug#15555)

2014-02-20  Glenn Morris  <rgm@gnu.org>

	* data.c (Fdefalias): Doc fix.

2014-02-19  Eli Zaretskii  <eliz@gnu.org>

	* xdisp.c (display_line): Fix horizontal scrolling of large images
	when fringes are turned off.  This comes at a price of not
	displaying the truncation/continuation glyphs in this case.
	(Bug#16806)

	* image.c (x_create_x_image_and_pixmap) [HAVE_NTGUI]: If
	CreateDIBSection returns an error indication, zero out *ximg after
	destroying the image.  This avoids crashes in memory allocations
	due to the fact that some of the callers also call
	x_destroy_x_image, which will attempt to free an already free'd
	block of memory.

2014-02-18  Martin Rudalics  <rudalics@gmx.at>

	* widget.c (update_various_frame_slots): Don't set
	FRAME_PIXEL_HEIGHT and FRAME_PIXEL_WIDTH here (Bug#16736).

2014-02-18  Michael Albinus  <michael.albinus@gmx.de>

	* dbusbind.c (xd_close_bus): Apply proper check on busobj.

2014-02-17  Paul Eggert  <eggert@cs.ucla.edu>

	temacs --daemon fix (Bug#16599).
	* emacs.c (main): Initialize daemon_pipe[1] here ...
	(syms_of_emacs): ... instead of here.

2014-02-16  Anders Lindgern  <andlind@gmail.com>

	* nsterm.m (keyDown:): Check for normal key even if NSNumericPadKeyMask
	is set (Bug#16505).

2014-02-16  Daniel Colascione  <dancol@dancol.org>

	* dbusbind.c (xd_lisp_dbus_to_dbus): New function.
	(xd_get_connection_address): Use it.
	(xd_close_bus): Use xd_lisp_dbus_to_dbus to instead of
	xd_get_connection_address because the latter signals if the bus
	we're trying to close is already disconnected.

2014-02-13  Eli Zaretskii  <eliz@gnu.org>

	* w32proc.c (start_timer_thread): Pass a non-NULL pointer as last
	argument to CreateThread.  This avoids segfaults on Windows 9X.
	Reported by oslsachem <oslsachem@gmail.com>.

2014-02-13  Paul Eggert  <eggert@cs.ucla.edu>

	Fix subspace record bug on HP-UX 10.20 (Bug#16717).
	* unexhp9k800.c (unexec_error): New function, to simplify the code.
	(check_lseek): New function, to report lseek errors.
	(save_data_space, update_file_ptrs, read_header, write_header)
	(copy_file, copy_rest, unexec): Use these news functions.
	(update_file_ptrs): Don't assume wraparound behavior when
	converting a large size_t value to off_t.

2014-02-13  Dmitry Antipov  <dmantipov@yandex.ru>

	* composite.c (fill_gstring_header): Pass positions as C integers
	and move parameters checking to...
	(Fcomposition_get_gstring): ...this function.  Handle case when
	buffer positions are in reversed order and avoid crash (Bug#16739).
	Adjust docstring.
	* buffer.c (validate_region): Mention current buffer in error message.

2014-02-12  Marcus Karlsson  <mk@acc.umu.se>  (tiny change)

	* image.c (pbm_load): Set to NO_PIXMAP on error (Bug#16683).

2014-02-12  Lars Ingebrigtsen  <larsi@gnus.org>

	* buffer.c (syms_of_buffer): Doc clarification (bug#9981).

2014-02-11  Glenn Morris  <rgm@gnu.org>

	* nsfns.m (ns_display_info_for_name, Fx_open_connection):
	Replace refs to "OpenStep" in messages.

2014-02-10  Paul Eggert  <eggert@cs.ucla.edu>

	Avoid "." at end of error diagnostics.
	* cmds.c (Fself_insert_command): Reword and avoid "." at end.
	* font.c (Ffont_at):
	* nsfns.m (ns_display_info_for_name):
	* nsselect.m (Fx_own_selection_internal):
	* nsterm.m (performDragOperation:):
	Remove "." from end of diagnostic.

2014-02-10  Lars Ingebrigtsen  <larsi@gnus.org>

	* fns.c (Fmaphash): Say what `maphash' returns, since it may be
	unintuitive (bug#15824).
	(Fyes_or_no_p): Doc fix (bug#15456).

2014-02-10  Dmitry Antipov  <dmantipov@yandex.ru>

	* cmds.c (Fself_insert_command): Respect the width of EMACS_INT
	and avoid warning.
	* eval.c (call_debugger): When exiting the debugger, do not allow
	max_specpdl_size less than actual binding depth (Bug#16603).
	(syms_of_eval): Adjust docstring.

2014-02-09  Lars Ingebrigtsen  <larsi@gnus.org>

	* cmds.c (Fself_insert_command): Output a clearer error message on
	negative repetitions (bug#9476).

	* macros.c (Fexecute_kbd_macro): Doc fix (bug#14206).

2014-02-08  Lars Ingebrigtsen  <larsi@gnus.org>

	* syntax.c (Fskip_syntax_backward): Doc clarification (bug#15115).

	* minibuf.c (Fread_string): Doc clarification (bug#15422).

	* buffer.c (Fmake_overlay): Doc clarification (bug#15489).

2014-02-08  Juanma Barranquero  <lekktu@gmail.com>

	* keyboard.c (Frecursive_edit): Fix typo in docstring.

2014-02-08  Lars Ingebrigtsen  <larsi@gnus.org>

	* xdisp.c (syms_of_xdisp): Doc clarification (bug#15657).

	* keyboard.c (Frecursive_edit): Say more precicely how throwing
	`exit' works (bug#15865).

2014-02-07  Martin Rudalics  <rudalics@gmx.at>

	Constrain window box/body sizes and margin widths (Bug#16649).
	* xdisp.c (window_box_width): Don't return less than zero.
	(window_box_left_offset, window_box_right_offset): Don't return
	more than the window's pixel width.
	* window.c (window_body_height, window_body_width): Don't return
	negative value.
	(window_resize_apply): Adjust margin width, if necessary.

2014-02-07  Glenn Morris  <rgm@gnu.org>

	* nsterm.m (syms_of_nsterm): Doc fix.

2014-02-06  Eli Zaretskii  <eliz@gnu.org>

	* w32.c (pMultiByteToWideChar, pWideCharToMultiByte):
	New variables: pointers through which to call the respective APIs.
	(filename_to_utf16, filename_from_utf16, filename_to_ansi)
	(filename_from_ansi, sys_link, check_windows_init_file):
	Call MultiByteToWideChar and WideCharToMultiByte through pointers.
	This is required on Windows 9X, where we dynamically load
	UNICOWS.DLL which has their non-stub implementations.
	(maybe_load_unicows_dll): Assign addresses to these 2 function
	pointers after loading UNICOWS.DLL.

	* w32fns.c (Fx_file_dialog, Fw32_shell_execute) [!CYGWIN]: Call
	MultiByteToWideChar and WideCharToMultiByte through function
	pointers.

	* w32.h (pMultiByteToWideChar, pWideCharToMultiByte):
	New declarations.

2014-02-06  Jan Djärv  <jan.h.d@swipnet.se>

	* nsterm.m (toggleFullScreen:): Hide menubar on secondary monitor
	for OSX >= 10.9 if separate spaces are used.
	(toggleFullScreen:): Use screen of w instead of fw (Bug#16659).

2014-02-06  Glenn Morris  <rgm@gnu.org>

	* buffer.c (cache-long-scans): Doc fix.

2014-02-05  Eli Zaretskii  <eliz@gnu.org>

	* w32fns.c (Fw32_shell_execute): Doc fix.

2014-02-05  Bastien Guerry  <bzg@gnu.org>

	* syntax.c (Fforward_word): Call Fconstrain_to_field with
	ESCAPE-FROM-EDGE set to `nil' (Bug#16453).

2014-02-05  Martin Rudalics  <rudalics@gmx.at>

	* fringe.c (draw_fringe_bitmap_1): Don't draw a fringe if it's
	outside the window (Bug#16649).

	* xdisp.c (note_mouse_highlight): When entering a margin area show
	a non-text cursor (Bug#16647).

2014-02-04  Paul Eggert  <eggert@cs.ucla.edu>

	* menu.c (Fx_popup_dialog): Remove label 'dialog_via_menu'.
	It prompted a compile-time diagnostic on GNU/Linux.
	Simplify to remove the need for the label.

2014-02-04  Eli Zaretskii  <eliz@gnu.org>

	* w32menu.c (w32_popup_dialog): Don't condition the whole function
	on HAVE_DIALOGS.  If the dialog is "simple", pop up a message box
	to show it; otherwise return 'unsupported--w32-dialog' to signal
	to the caller that emulation with menus is necessary.
	This resurrects code inadvertently deleted by the 2013-10-08 commit.
	(Bug#16636)
	(syms_of_w32menu): DEFSYM Qunsupported__w32_dialog.

	* w32term.h (w32_popup_dialog): Prototype is no longer conditioned
	by HAVE_DIALOGS.

	* menu.c (Fx_popup_dialog): Don't condition the call to
	w32_popup_dialog on HAVE_DIALOGS.  If w32_popup_dialog returns a
	special symbol 'unsupported--w32-dialog', emulate the dialog with
	a menu by calling x-popup-menu.

	* menu.h (Qunsupported__w32_dialog): New extern variable.

2014-02-04  Michael Albinus  <michael.albinus@gmx.de>

	* keyboard.c (kbd_buffer_get_event): Read file notification events
	also in batch mode.

	* xgselect.c (xg_select): Read glib events in any case, even if
	there are no file descriptors to watch for.  (Bug#16519)

2014-02-03  Martin Rudalics  <rudalics@gmx.at>

	* dispextern.h (face_id): Add WINDOW_DIVIDER_FIRST_PIXEL_FACE_ID
	and WINDOW_DIVIDER_LAST_PIXEL_FACE_ID.
	* w32term.c (w32_draw_window_divider): Handle first and last
	pixels specially.
	* w32term.h (w32_fill_area_abs): New function.
	* xdisp.c (x_draw_right_divider): Don't draw over bottom
	divider.
	* xfaces.c (realize_basic_faces): Handle new face ids.
	* xfns.c (Fx_create_frame): Call x_default_parameter for right
	and bottom divider width.
	* xterm.c (x_draw_window_divider): Handle first and last pixels
	specially.

2014-02-03  Dmitry Antipov  <dmantipov@yandex.ru>

	* print.c (Fexternal_debugging_output): Add cast to pacify
	--enable-gcc-warnings.
	* eval.c (call_debugger): Grow specpdl if the debugger was
	entered due to specpdl overflow (Bug#16603) and allow more
	specpdl space for the debugger itself.

2014-02-02  Martin Rudalics  <rudalics@gmx.at>

	* w32fns.c (Fx_create_frame): Process frame alpha earlier.
	(Bug#16619)

2014-02-01  Eli Zaretskii  <eliz@gnu.org>

	* w32fns.c (Ffile_system_info): Use WINAPI in the function
	pointers that get the address of GetDiskFreeSpaceEx.  (Bug#16615)

	* print.c (Fexternal_debugging_output): If the argument character
	is non-ASCII, encode it with the current locale's encoding before
	writing the result to the terminal.  (Bug#16448)

	* w32fns.c (Fw32_shell_execute): Don't call file-exists-p for
	DOCUMENT that is a "remote" file name, i.e. a file-handler exists
	for it.  (Bug#16558)

2014-01-30  Andreas Schwab  <schwab@linux-m68k.org>

	* process.c (create_process): Reset SIGPROF handler in the child.
	* callproc.c (call_process): Likewise.

2014-01-29  Paul Eggert  <eggert@cs.ucla.edu>

	* xmenu.c (create_and_show_popup_menu): Port comment to C89.

2014-01-29  Eli Zaretskii  <eliz@gnu.org>

	* .gdbinit (xprintstr, xprintbytestr): Don't use repetition count
	of zero to print strings, GDB doesn't like it.

	* print.c (print_object): Use FETCH_STRING_CHAR_ADVANCE, not
	STRING_CHAR_AND_LENGTH, so that if the string is relocated by GC,
	we still use correct addresses.  (Bug#16576)

2014-01-27  K. Handa  <handa@gnu.org>

	Fix bug#16286 by a different method from 2014-01-26T00:32:30Z!eggert@cs.ucla.edu,
	to preserve the code detection behavior of 24.3.
	* coding.h (struct coding_system): New member detected_utf8_bytes.
	* coding.c (detect_coding_utf_8): Set coding->detected_utf8_bytes.
	(decode_coding_gap): Use short cut for UTF-8 file reading only
	when coding->detected_utf8_bytes equals to coding->src_bytes.
	* fileio.c (Finsert_file_contents): Cancel the previous change.

2014-01-29  Martin Rudalics  <rudalics@gmx.at>

	* w32fns.c (x_set_tool_bar_lines): Don't clear area on frames
	that are not visible.

2014-01-29  Jan Djärv  <jan.h.d@swipnet.se>

	* xmenu.c (create_and_show_popup_menu): Handle case when no key
	is grabbed (Bug#16565).

2014-01-28  Martin Rudalics  <rudalics@gmx.at>

	* xdisp.c (last_max_ascent): Re-remove after erroneously
	reintroducing it on 2013-11-30 and abolishing Dmitry's removal
	from 2013-03-29.
	(move_it_to): Re-remove reference to last_max_ascent.
	(Fwindow_text_pixel_size): Add iterator's max_ascent and
	max_descent here instead of calling line_bottom_y.
	Fix doc-string.

2014-01-28  Dmitry Antipov  <dmantipov@yandex.ru>

	* terminal.c (initial_free_frame_resources): New function.
	(init_initial_terminal): Install new hook to free face cache
	on initial frame and avoid memory leak.  For details, see
	<http://lists.gnu.org/archive/html/emacs-devel/2014-01/msg01974.html>.
	* xfaces.c (free_frame_faces): Adjust comment.

2014-01-26  Paul Eggert  <eggert@cs.ucla.edu>

	* data.c (Fstring_to_number): Document results if unparsable
	(Bug#16551).

2014-01-26  Jan Djärv  <jan.h.d@swipnet.se>

	* xterm.c (x_focus_changed): Check for non-X terminal-frame (Bug#16540)

2014-01-26  Paul Eggert  <eggert@cs.ucla.edu>

	When decoding, prefer ptrdiff_t to int for buffer positions etc.
	* coding.c (detect_coding_utf_8, emacs_mule_char)
	(detect_coding_iso_2022, encode_coding_iso_2022, check_ascii)
	(check_utf_8, decode_coding):
	* coding.h (struct coding_system.errors):
	Use ptrdiff_t, not int, for integer values derived from buffer and
	string positions.

	Fix crash with insert-file-contents and misdecoded text (Bug#16286).
	* fileio.c (Finsert_file_contents): Set CODING_MODE_LAST_BLOCK
	before invoking decode_coding_gap, since there's just one block.

2014-01-25  Martin Rudalics  <rudalics@gmx.at>

	Fix handling of face attributes in Fx_create_frame (Bug#16529).
	* w32fns.c (Fx_create_frame): Don't inhibit running Lisp code
	too early.  Again run change_frame_size before assigning menu-
	and tool-bar-lines.

2014-01-25  Fabrice Popineau  <fabrice.popineau@gmail.com>

	* w32term.c (w32_read_socket): When the WM_DISPLAYCHANGE message
	arrives, call x_check_fullscreen, in case the new display has a
	different resolution.  (Bug#16517)