Reinstate some of the functionality removed in commit
67a878f78f.
* src/nsimage.m (ns_image_from_XBM): Use new reverseBytes argument.
([EmacsImage initFromXBM:width:height:fg:bg:reverseBytes:]): Add
ability to reverse the contents of each byte for use with XBMs, while
still working with fringe bitmaps.
* src/nsterm.h
([EmacsImage initFromXBM:width:height:fg:bg:reverseBytes:]): Modified
function definition.
* src/nsterm.m (ns_draw_fringe_bitmap): Use new reverseBytes argument.
* src/minibuf.c (read_minibuf): Avoid restoring the previous buffer,
as this is already done at the end of the function; call
`add-to-history' after that point.
* src/keyboard.c (read_key_sequence): Modify the fix for
bug#37782 so that it applies only to keyboard keys, not to
other events, such as a new buffer. (Bug#38132)
* src/marker.c (set_marker_internal): Handle the case where
POSITION is beyond PTRDIFF_MAX, which can happen if Emacs was
built --with-wide-int. Bug uncovered by the recently added
overlay tests.
* lisp/simple.el (minibuffer-history-values): New function, should be
used to access the minibuffer input history variable when the
minibuffer might be active. If the variable is buffer-local, the
previous buffer's value will be used.
(goto-history-element): Use the new function to access the minibuffer
history.
(minibuffer-history-isearch-wrap): Use the new function to access the
minibuffer history.
* src/minibuf.c (read_minibuf): Switch to previous buffer temporarily
before updating history list (Bug#38317).
(read-from-minibuffer): Extend documentation to mention that the
result of using the command will be added to the history list by
default.
* doc/lispref/minibuf.texi (Minibuffer History): Mention the
possibility of using a buffer-local variable as history.
* etc/NEWS: Announce changes.
As described in the new comment added to emacs-module.c, using GMP
directly in the module interface has significant downsides: it couples
the module interface directly to the implementation and requires
module authors to link their module against the same GMP library as
Emacs itself, which is often difficult and an unnecessary burden. By
picking a representation for the magnitude that often matches the one
used by GMP, we can avoid overhead when converting from and to GMP in
most cases.
Loading the test module in test/data/emacs-module and evaluating
(dotimes (_ 10000)
(mod-test-double (* 2 most-negative-fixnum)))
under Callgrind shows that on my (GNU/Linux) machine Emacs only spends
10% of the CPU time of mod-test-double in mpz_import and mpz_export
combined, even though that function does little else. (By contrast,
30% is spent in allocate_pseudovector.)
* src/emacs-module.h.in: Don't check EMACS_MODULE_GMP. Don't include
gmp.h. Remove emacs_mpz structure. Instead, define type alias
emacs_limb_t and macro EMACS_LIMB_MAX.
* src/module-env-27.h: Change interface of extract_big_integer and
make_big_integer to take a sign-magnitude representation instead of
mpz_t.
* src/emacs-module.c: Don't check EMACS_MODULE_GMP or
EMACS_MODULE_HAVE_MPZ_T. Add a comment about the chosen
implementation.
(module_extract_big_integer, module_make_big_integer): Reimplement
without using mpz_t in the interface.
* doc/lispref/internals.texi (Module Values): Adapt function
documentation and example. Stop mentioning GMP and EMACS_MODULE_GMP.
* test/data/emacs-module/mod-test.c: Don't define EMACS_MODULE_GMP or
EMACS_MODULE_HAVE_MPZ_T.
(memory_full, extract_big_integer, make_big_integer): New helper
functions, identical to example in the Info documentation.
(Fmod_test_nanoseconds, Fmod_test_double): Adapt to new interface.
This should fix Bug#31223, Bug#28106, Bug#23672 as well as Ubuntu bug
https://bugs.launchpad.net/ubuntu/+source/emacs25/+bug/1695228
Also fixes the formerly unscaled Y value returned by
frame-monitor-workarea (and display-monitor-attributes-list).
For details on why some GTK menus were empty please see thread
https://lists.gnu.org/archive/html/emacs-devel/2019-11/msg01061.html
* src/gtkutil.c (menubar_map_cb, xg_update_frame_menubar): Scale up
req.height so that the menu bar's height is in device pixels as
expected.
(xg_event_is_for_menubar): Scale down rec.x and rec.y so that
gtk_widget_intersect() works as intended.
* src/xfns.c (Fx_display_monitor_attributes_list): Scale work.x and
work.y up to be in device pixels.
Copyright-paperwork-exempt: yes
* src/image.c (lookup_image): Move call to image_set_transform after
postprocess_image.
(image_create_x_image_and_pixmap_1): Use new function.
(image_set_transform): Apply the transform to the mask too.
(x_create_xrender_picture): New function.
(Create_Pixmap_From_Bitmap_Data):
(xpm_load): Use new function.
* src/xterm.c (x_composite_image): Use PictOpOver when there is a mask
so the transparency is honoured.
(x_draw_image_foreground_1): Use x_composite_image.
* src/xdisp.c (syms_of_xdisp): New variable 'bidi-inhibit-bpa'.
* src/bidi.c (bidi_paired_bracket_type): If 'bidi-inhibit-bpa'
is non-nil, return BIDI_BRACKET_NONE for all characters.
* lisp/frame.el: Add 'bidi-inhibit-bpa' to the list of
variables whose changes require redisplay. (Bug#38407)
Make it possible to select a rectangular region using the mouse.
The standard binding is C-M-mouse-1.
* lisp/mouse.el (mouse-scroll-subr): Add ADJUST argument.
(mouse-drag-region-rectangle): New.
* lisp/rect.el (rectangle--reset-point-crutches): New.
(rectangle--reset-crutches): Use 'rectangle--reset-point-crutches'.
* src/xdisp.c (remember_mouse_glyph, syms_of_xdisp):
Add 'mouse-fine-grained-tracking'.
* doc/lispref/commands.texi (Motion Events):
Document 'mouse-fine-grained-tracking'.
* doc/emacs/frames.texi (Mouse Commands):
* doc/emacs/killing.texi (Rectangles):
* etc/NEWS: Document rectangular selection with the mouse.
For the ns-KEY-modifier and ns-right-KEY-modifier variables, KEY
being 'control', 'command', 'alternate' and 'function', allow
values on the form (:ordinary SYMBOL :function :SYMBOL :mouse SYMBOL),
so that the key can be used for different modifiers (or none) in
different contexts. This is particularly useful for using the macOS
Option key for extended character entry while still using it as an
Emacs modifier for function keys and mouse clicks.
* src/nsterm.m (mod_of_kind, right_mod, nil_or_none): Helper functions.
(EV_MODIFIERS2): Add KIND argument.
(EV_MODIFIERS): Adapt call to EV_MODIFIERS2.
(ns_get_shifted_character): Use correct event kind for modifiers.
(ns-alternate-modifier, ns-right-alternate-modifier)
(ns-command-modifier, ns-right-command-modifier)
(ns-control-modifier, ns-right-control-modifier)
(ns-function-modifier): Rewrite doc strings for new data format.
(QCordinary, QCfunction, QCmouse): Define symbols.
* lisp/cus-start.el: Conform to new data types.
* doc/emacs/macos.texi (Mac / GNUstep Basics)
(Mac / GNUstep Customization): Improved documentation.
* etc/NEWS: Mention the change.
* doc/lispref/display.texi (Displaying Messages): Explain the
behavior of using minibuffer-message if the minibuffer is active.
* src/editfns.c (Fmessage_in_echo_area): New function with body
copied from Fmessage.
(Fmessage): Call minibuffer-message in the active minibuffer,
otherwise call Fmessage_in_echo_area.
(message-in-echo-area): New variable.
* lisp/isearch.el (isearch--momentary-message, isearch-message):
* lisp/minibuffer.el (minibuffer-message, minibuffer-completion-help):
Use 'message-in-echo-area' instead of 'message' where necessary.
* lisp/autorevert.el (auto-revert-handler):
* lisp/man.el (Man-bgproc-sentinel):
* lisp/subr.el (do-after-load-evaluation):
Revert recent changes that replaced 'message' with 'minibuffer-message'.
This is not needed anymore since 'message' uses 'minibuffer-message'
in the active minibuffer.
* src/xfaces.c: Update and improve commentary at the beginning
of the file.
(face_attr_sym): New static array.
(init_xfaces): Initialize 'face_attr_sym'.
(merge_face_ref): Handle the :extend attribute in faces
specified as lists of key/value pairs. (Bug#37774)
mingw.org's MinGW by default targets Windows 9X, so
_WIN32_WINNT is set to a value that bypasses declarations
in system headers we need to compile network_interface_list.
Also, the code needed a workaround for Windows XP, where
some functionality is missing from the GetAdaptersAddresses
API.
* src/w32.c (_WIN32_WINNT): Define to 0x0501, if the value is
lower, temporarily while processing iphlpapi.h.
(address_prefix_match): New helper function.
(network_interface_list): Work around the fact that the
OnLinkPrefixLength member of IP_ADAPTER_UNICAST_ADDRESS is not
available when _WIN32_WINNT < 0x0600. On Windows XP use
special code that calls address_prefix_match to compute the
network prefix length.
Bug#38218
* src/process.c (Fnetwork_interface_list): Extend argument list to
allow requesting full network info and/or IPv4/IPv6 info.
(network_interface_list) [HAVE_GETIFADDRS]: Use getifaddrs to retrieve
interface IP addresses.
* src/process.h: Update prototype of network_interface_list.
* src/w32.c (g_b_init_get_adapters_addresses): New init flag.
(globals_of_w32): Initialize it.
(GetAdaptersAddresses_Proc): New function typedef.
(get_adapters_addresses): New wrapper function.
(init_winsock): Load htonl and ntohl.
(sys_htonl, sys_ntohl): New wrapper functions.
(network_interface_list): Implement in terms of
get_adapters_addresses.
* nt/inc/sys/socket.h: Add sys_htonl and sys_ntohl prototypes.
* etc/NEWS: Announce IPv4/IPv6 changes in network-interface-list.
* doc/lispref/processes.texi (Misc Network): Document updated arglist
and return values for network-interface-list.
* src/xfaces.c (face_inherited_attr): New function.
(merge_named_face): Call 'face_inherited_attr' when testing
whether a face that inherits from another fits the filtering
criteria specified by ATTR_FILTER.
(merge_face_vectors): Revert the changes made in this function
for filtering by ATTR_FILTER, and remove that argument as
well. These tests are now completely done by the caller, see
'merge_named_face'. (Bug#37774)
* src/xdisp.c (append_space_for_newline): Add an eassert check
that default_face is not null, by calling FACE_FROM_ID instead of
FACE_FROM_ID_OR_NULL. Initialize a local only if needed.
* src/gnutls.c: Move the "#ifdef HAVE_GNUTLS" earlier, so
that "./configure --enable-gcc-warnings --without-gnutls"
does not complain about macros being defined but never used.
Indent "#" directives more consistently.
* src/xdisp.c (extend_face_to_end_of_line): Make sure the
character position of the stretch glyph inserted to extend the
face is zero, as various other parts of the display code rely
on that. (Bug#38330)
* src/coding.c (get_char_bytes, encode_string_utf_8)
(decode_string_utf_8): Fix commentary.
(encode_string_utf_8): Return the original ASCII string only
if NOCOPY is non-zero.
(decode_string_utf_8): Accept 2 additional arguments STR and
STR_LEN, which allow to pass the input text as a C string.
(make_string_from_utf8): Delegate the job to decode_string_utf_8.
* src/coding.h: Update the prototype of decode_string_utf_8.
* src/json.c (json_encode): Call encode_string_utf_8.
Remember the full font description instead of just the family so that
size/style/weight settings are preserved.
* gtkutil.c (xg_get_font) [HAVE_GTK3]: Use the pango font
description to set/get the current font (Bug#28901).
Our minimum GTK3 version is 3.10, the font filter functions appeared
in 3.2.
* gtkutil.c (xg_font_filter) [HAVE_GTK3]: Just check for HAVE_GTK3.
(xg_get_font) [HAVE_GTK3]: Same here.
* src/xftfont.c (xftfont_open): FC_CHARCELL is apparently an alias
for FC_DUAL used in some east Asian fonts (bug#35079). Modelled
after a patch suggested by Kenichi Handa.
* src/w32font.c (w32_to_fc_weight): Support a few more weight
values, for compatibility with the GTK font selection widget
(see gtkutil.c:XG_WEIGHT_TO_SYMBOL). (Bug#24226)
The old code assumed that hexdigit initialization (needed by
non-GCC) could be done in syms_of_character, but that is no longer
true with pdumper. Instead, simplify hexdigit init so that it can
be done statically on all C99 platforms. Problem discovered on
Solaris 10 sparc + Oracle Solaris Studio 12.6.
* src/character.c (hexdigit): Add 1 to every value; all uses
changed. This simplifies the initialization so that it can be
done statically on any C99 compiler. hexdigit is now always const.
(syms_of_character): Omit no-longer-necessary initialization.
* src/character.h (HEXDIGIT_CONST, HEXDIGIT_IS_CONST):
Remove. All uses removed.
* doc/lispref/variables.texi (Watching Variables): Clarify the
documentation of 'add-variable-watcher' and fix markup.
* src/data.c (Fadd_variable_watcher): Clarify the doc string.
(Bug#38205)
* src/xdisp.c (tty_handle_tab_bar_click): Revert the last
change which made this work only in the HAVE_NTGUI build.
This function is needed by any build which supports a mouse
on TTY frames.
* src/floatfns.c (double_integer_scale): Distinguish Inf from NaN.
(rounding_driver): Handle (ceil 0 0.0) and (ceil 0 1.0e+INF).
* test/src/floatfns-tests.el (special-round): Add tests for
weird cases like this.
Avoid crash with (floor 0 0.0)
* src/floatfns.c (rounding_driver): Signal an arithmetic
error if divisor is 0.0 or -0.0, instead of crashing.
