This should fix several bugs reported recently against Emacs
(Bug#28518, Bug#28626, Bug#28632, Bug#28633, Bug#28770).
* src/image.c (imagemagick_load_image):
Do not call MagickWandTerminus.
* src/indent.c (line_number_display_width): Make sure a line
number is generated even when the first line of the window is
completely hscrolled out of view. (Bug#28854)
Problem reported by Roland Winkler (Bug#28824#35).
Based on a patch proposed by Andy Moreton (Bug#28824#38).
* src/image.c (pbm_scan_index): New function.
(pbm_load): Use it to decode raw data correctly when its top bit
is set.
Make the effect of customizing arabic-shaper-ZWNJ-handling appear
immediately.
* lisp/language/misc-lang.el (arabic-shaper-ZWNJ-handling): Add :set
value.
* src/composite.c (Fclear_composition_cache): New function.
(syms_of_composite): Defsubr it.
* src/systhread.c: Include w32term.h instead of windows.h.
(w32_beginthread_wrapper): Add ALIGN_STACK attribute, to ensure
the thread's stack is correctly aligned.
* src/w32term.h (ALIGN_STACK): Update commentary.
* src/gnutls.c (HAVE_GNUTLS_MAC_GET_NONCE_SIZE)
[GNUTLS_VERSION_NUMBER >= 0x030300]: Define.
(gnutls_mac_get_nonce_size): Depend on
HAVE_GNUTLS_MAC_GET_NONCE_SIZE.
(Fgnutls_macs): Use 0 for :mac-algorithm-noncesize if
!HAVE_GNUTLS_MAC_GET_NONCE_SIZE.
Problem diagnosed by Andreas Schwab (Bug#28792#65).
This fixes a bug that I introduced in
2017-09-10T22:39:24@eggert@cs.ucla.edu
"Fix race with rename-file etc. with dir NEWNAME".
* src/fileio.c (Frename_file): Copy a source directory across
file system boundaries even if its name does not end in slash.
* src/sysdep.c (get_current_dir_name_or_unreachable):
Avoid integer overflow if working directory name is absurdly long.
When allocating memory for getcwd, do not exceed MAXPATHLEN.
* src/xdisp.c (redisplay_window): Clear the window's desired glyph
matrix before calling try_window with the same starting point.
This avoids assertion violations when switching from a buffer
without line numbers to a buffer with line numbers. (Bug28710)
* src/sysdep.c (get_current_dir_name_or_unreachable):
New function, with most of the old contents of
emacs_get_current_dir_name.
(emacs_get_current_dir_name): Use it. Use a simpler
test for unreachable directory strings, and also apply
it to getcwd etc. (Bug#27871)
* src/dispnew.c (update_frame_line): Accept an additional argument
UPDATING_MENU_P; if non-zero, home the cursor before updating a
frame's line. All callers changed.
(update_frame_1): Accept an additional argument UPDATING_MENU_P,
and pass it to update_frame_line. All callers changed.
(update_frame_with_menu): Call update_frame_1 with last argument
non-zero. (Bug#17497)
Problem reported by Kaushal Modi in:
http://lists.gnu.org/archive/html/emacs-devel/2017-10/msg00141.html
* src/editfns.c (styled_format): Fix bug where USE_SAFE_ALLOCA was
not always followed by SAFE_FREE. This bug was introduced in my
patch 2017-09-26T23:31:57Z!eggert@cs.ucla.edu entitled "Avoid some
unnecessary copying in Fformat etc."
Although the Lisp manual said that ‘format’ returns a
newly-allocated string, this was not true for a few cases like
(format "%s" ""), and fixing the documentation to allow reuse of
arguments lets us improve performance in common cases like
(format "foo") and (format "%s" "foo") (Bug#28625).
* doc/lispref/strings.texi (Formatting Strings):
* etc/NEWS:
Say that the result of ‘format’ might not be newly allocated.
* src/callint.c (Fcall_interactively):
* src/dbusbind.c (XD_OBJECT_TO_STRING):
* src/editfns.c (Fmessage, Fmessage_box):
* src/xdisp.c (vadd_to_log, Ftrace_to_stderr):
Just use Fformat or Fformat_message, as that’s simpler and no
longer makes unnecessary copies.
* src/editfns.c (styled_format): Remove last argument, as it
is no longer needed: all callers now want it to behave as if it
were true. All remaining callers changed. Make this function
static again. Simplify the function now that we no longer
need to worry about whether the optimization is allowed.
* src/thread.h (m_getcjmp): New member of 'struct thread_state'.
(getcjmp): Define to current thread's 'm_getcjmp'.
* src/thread.c (maybe_reacquire_global_lock): Switch to main
thread, since this is called from a SIGINT handler, which always
runs in the context of the main thread.
* src/lisp.h (sys_jmp_buf, sys_setjmp, sys_longjmp): Move the
definitions before thread.h is included, as thread.h now uses
sys_jmp_buf.
* src/keyboard.c (getcjmp): Remove declaration.
(read_char): Don't call maybe_reacquire_global_lock here.
(handle_interrupt): Call maybe_reacquire_global_lock here, if
invoked from the SIGINT handler, to make sure
quit_throw_to_read_char runs with main thread's Lisp bindings and
uses the main thread's jmp_buf buffer. (Bug#28630)
* src/frame.c (Ficonify_frame) [HAVE_WINDOW_SYSTEM]: Use
frame_parent only in GUI builds to avoid compilation errors in
--without-x builds. (Bug#28611)
This fixes some URLs I omitted from my previous pass,
notably those in lists.gnu.org. Although lists.gnu.org
does not yet support TLS 1.1, TLS 1.0 is better than nothing.
* lisp/erc/erc.el (erc-official-location):
* lisp/mail/emacsbug.el (report-emacs-bug):
Use https:, not http:.
This is a partial reversion of yesterday's commit by the same author, which
changed the meaning of nil and introduced the new value t.
* src/doc.c (text_quoting_style, text-quoting-style)
(internal--text-quoting-flag): Revert yesterday's changes.
* lisp/cus-start.el: (top level): Amend the entry for text-quoting-style.
* etc/NEWS: Amend the entry for text-quoting-style.
* doc/lispref/control.texi (Signalling Errors)
* doc/lispref/display.texi (Displaying Messages)
* doc/lispref/strings.texi (Formatting Strings): Bind text-quoting-style to
grave rather than nil to inhibit translation of quotes.
* doc/lispref/help.texi (Keys in Documentation): Revert the description of the
proposed new default, t.
A value of nil for text-quoting-style now means "no translation". t means
"Use curved quotes if displayable".
* src/doc.c (text-quoting-style (function)): modify for new semantics.
(text-quoting-style (variable)): Amend the doc string, set the default value
to t.
* lisp/cus-start.el: (top level): Create a customize entry for
text-quoting-style in group display.
* etc/NEWS: Amend the entry for text-quoting-style.
* doc/emacs/display.texi (Text Display): Describe the translation of ASCII
quotes to curved quotes, and how to influence or inhibit it.
* doc/lispref/control.texi (Signalling Errors)
* doc/lispref/display.texi (Displaying Messages)
* doc/lispref/strings.texi (Formatting Strings): Describe binding
text-quoting-style to nil to inhibit unwanted quote translation.
* doc/lispref/help.texi (Keys in Documentation): Change text-quoting-style
from a variable to a user option. Describe its changed set of values. State
that it can be customized freely.
* src/w32term.c (syms_of_w32term) [x-wait-for-event-timeout]: New
variable.
(x_make_frame_visible): Wait for frame to become visible according to
its value.
(input_signal_count): Remove.
But wait specfically for a MapNotify event, and only for a
configurable amount of time.
* src/xterm.c (syms_of_xterm) [x-wait-for-event-timeout]: New
variable.
(x_wait_for_event): Use it instead of hardcoding the wait to 0.1s.
(x_make_frame_visible): Call x_wait_for_event at the end.
* etc/NEWS: Announce x_wait_for_event.
This patch is just for performance; it should not affect behavior.
On my platform, it made the microbenchmark (format "%S" load-path)
run about 45% faster. It should also speed up calls like (message
"%s" STRING).
* src/callint.c (Fcall_interactively):
* src/dbusbind.c (XD_OBJECT_TO_STRING):
* src/editfns.c (Fmessage, Fmessage_box):
* src/xdisp.c (vadd_to_log, Ftrace_to_stderr):
Use styled_format instead of Fformat or Fformat_message,
to avoid unnecessary copying.
* src/editfns.c (styled_format): New arg NEW_RESULT.
All uses changed. Reuse an input string if it has the
right value and if !NEW_RESULT.
* src/lisp.h (style_format): New decl.
* src/nsterm.m (EmacsView::updateFrameSize): Don't wait for the
toolbar on undecorated frames.
(EmacsView::initFrameFromEmacs): Group window flags correctly.
