src/dispextern.h (struct bidi_it): New member next_en_type.
src/bidi.c (bidi_line_init): Initialize the next_en_type member.
(bidi_resolve_explicit_1): When next_en_pos is valid for the
current character, check also for next_en_type being WEAK_EN.
(bidi_resolve_weak): Don't enter the expensive loop if the current
position is before next_en_pos. Record the bidi type of the first
non-ET, non-BN character we find, in addition to its position.
(bidi_level_of_next_char): Invalidate next_en_type when
next_en_pos is over-stepped.
* configure.in (AC_STRUCT_TM, AC_STRUCT_TIMEZONE, HAVE_TM_GMTOFF):
Remove; no longer needed, now that we defer to strftime for time
zone names.
* src/editfns.c: Rewrite current-time-zone so that it invokes
the equivalent of (format-time-string "%Z") to get the time zone name.
This fixes a bug when the time zone name contains characters that
need converting from the system time locale to Emacs internal format.
This fixes a shortcoming that I introduced in my 1999-10-19 patch:
that patch fixed format-time-string to do the conversion, but
I forgot to fix current-time-zone.
(format_time_string): New function, containing most of
what Fformat_time_string used to contain.
(Fformat_time_string): Rewrite in terms of format_time_string.
This doesn't change this function's behavior.
(current-time-zone): Rewrite to use format_time_string.
This fixes the bug reported by Michael Schierl in
<http://lists.gnu.org/archive/html/emacs-devel/2007-06/msg00334.html>.
Jason Rumney's 2007-06-07 change worked around this bug, but
didn't fix it.
* src/systime.h (tzname, timezone): Remove no-longer-used declarations.
src/xdisp.c (try_window_reusing_current_matrix): If a line ends in a display
vector or the next line starts in a display vector, continue
redrawing the window even though the character position of
start_row was reached.
Fixes: debbugs:9771
src/bidi.c (bidi_line_init): Initialize next_en_pos to zero, not -1.
(bidi_resolve_neutral): Don't enter the expensive loop looking for
non-neutral characters if the current character is a paragraph
separator (a.k.a. Newline). This avoids running the same
expensive loop twice, once when we consume the preceding newline
and the other time when the line actually needs to be displayed.
Avoid the loop when we see neutrals on the base embedding level
following a character whose directionality is the same as the
paragraph's. This avoids running the expensive loop when a line
ends in a long sequence of neutrals, like control characters.
Add assertion against STRONG_AL type. Slightly rearrange code
that determines the type of a neutral given the first non-neutral
that follows it.
(bidi_level_of_next_char): Set next_en_pos to zero when
invalidating its info.
for a line that begins with control characters.
src/xdisp.c (push_display_prop): Determine whether to record string
or buffer position by IT->string, not by IT->method. Allow
GET_FROM_DISPLAY_VECTOR as IT->method on entry.
src/bidi.c (bidi_init_it): Don't initialize paragraph_dir here, as
some callers set that themselves.
src/xdisp.c (Fcurrent_bidi_paragraph_direction): Initialize
itb.paragraph_dir to NEUTRAL_DIR. Fixes an occasional incorrect
cursor motion with <left> and <right> arrow keys.
* configure.in (GC_LISP_OBJECT_ALIGNMENT): Remove.
This is now done by src/alloc.c.
* src/alloc.c (GC_LISP_OBJECT_ALIGNMENT): Use offsetof, not __alignof__
or sizeof. __alignof__ gives the wrong answer on Fedora x86-64
with GCC 4.6.1 when configured with CC='gcc -m32' --with-wide-int;
this makes Emacs dump core during garbage collection on rare
occasions. sizeof is obviously inferior to offsetof here, so
stick with offsetof.
(GC_POINTER_ALIGNMENT): New macro.
(mark_memory): Omit 3rd (offset) arg; caller changed.
Don't assume EMACS_INT alignment is the same as pointer alignment.
(pcomplete--common-suffix): Remove.
(pcomplete--common-quoted-suffix): Use comint--common-suffix instead.
(pcomplete--table-subvert): Sync with comint--table-subvert.
(pcomplete--entries): Use comint-completion-file-name-table.
* lisp/comint.el (comint-unquote-filename): Simplify.
(comint-completion-file-name-table): New function.
(comint--complete-file-name-data): Use it.
* src/dired.c (file_name_completion): Don't expand file name.
(Ffile_name_completion, Ffile_name_all_completions): Expand file name
before checking file name handler.
Fixes: debbugs:9616
This fixes a Y2038 bug on 64-bit hosts.
* buffer.c (reset_buffer):
* fileio.c (Fdo_auto_save, Fset_buffer_auto_saved)
(Fclear_buffer_auto_save_failure):
Use 0, not -1, to represent an unset failure time, since time_t
might not be signed.
* alloc.c (XMALLOC_OVERRUN_CHECK_OVERHEAD, XMALLOC_OVERRUN_CHECK_SIZE):
Move back here from lisp.h, but with their new implementations.
(XMALLOC_BASE_ALIGNMENT, COMMON_MULTIPLE, XMALLOC_HEADER_ALIGNMENT)
(XMALLOC_OVERRUN_SIZE_SIZE): Move these new lisp.h macros here.
* charset.c (charset_table_init): New static var.
(syms_of_charset): Use it instead of xmalloc. This removes a
dependency on glibc malloc internals. See Eli Zaretskii's comment in
<http://lists.gnu.org/archive/html/emacs-devel/2011-09/msg00815.html>.
* lisp.h (XMALLOC_OVERRUN_CHECK_OVERHEAD, XMALLOC_OVERRUN_CHECK_SIZE):
Move back to alloc.c.
(XMALLOC_BASE_ALIGNMENT, COMMON_MULTIPLE, XMALLOC_HEADER_ALIGNMENT)
(XMALLOC_OVERRUN_SIZE_SIZE): Move to alloc.c.
