This can happen when running emacs over a forwarded X11 connection.
* src/pgtkfns.c (pgtk_get_monitor_scale_factor): Return 0 if we're
passed a NULL model name.
* src/pgtkterm.c (pgtk_term_init): Make reading GNOME settings optional.
(pgtk_text_scaling_factor): New function to read text scaling factor.
Copyright-paperwork-exempt: yes
* src/editfns.c (styled_format): Fix accounting for text
properties that come from the format string. (Bug#46317)
* test/src/editfns-tests.el (format-properties): Add new tests for
bug#46317.
* src/xdisp.c (Fwindow_text_pixel_size): Fix last change: preserve
the original Y coordinate after start_display, instead of zeroing
it out. Reported by martin rudalics <rudalics@gmx.at>.
* src/lread.c (Fread_char, Fread_event, Fread_char_exclusive):
Call cancel_echoing to make sure the prompt is not obscured by
keystrokes echo. (Bug#46243)
* src/xdisp.c (Fwindow_text_pixel_size): Support the use case
where FROM and TO belong to the same screen line. Reported by
Yuan Fu <casouri@gmail.com>.
* src/lread.c (invalid_syntax_lisp): Instead of putting the
line/column in a string, signal an error containing the numbers as
data. This allows for easier post-processing and is how other
similar errors (like (forward-sexp 1)) do it.
* src/lisp.h: Add count_lines prototype.
* src/lread.c (invalid_syntax_lisp): New function (bug#36970).
(invalid_syntax): Extend function to take a readcharfun parameter.
(read_emacs_mule_char, character_name_to_code): Pass in.
(read_escape, invalid_radix_integer, read1): Ditto.
* src/xdisp.c (count_lines): Add a more succinct shim over
display_count_lines.
src/minibuf.c (get_minibuffer): Move the XSETCAR which writes the new
minibuffer into Vminibuffer_list to immediately after the MB's creation, so
that the list is in a consistent state before calling fundamental-mode or
minibuffer-inactive-mode.
...when enable-recursive-minibuffers is non-nil, and several minibuffers are
activated from different frames. Also set the major mode of a reused active
minibuffer to `fundamental-mode' - up till now it's been
minibuffer-inactive-mode.
* src/minibuf.c (read_minibuf): with the indicated settings of variables,
"stack up" all containing minibuffers on the mini-window of the current
frame. Delete another, now superfluous such stacking up.
(set_minibuffer_mode): New function.
(get_minibuffer): Call the above new function (twice), in place of inline
code, ensuring active minibuffers are never left in minibuffer-inactive-mode.
The existing XIC implementation doesn't work when using Gtk IM Context.
Instead, `gtk_im_context_set_cursor_location` should be used
to update the cursor information for the input method.
* src/pgtkim.c (pgtk_im_set_cursor_location): New function to tell gtk
cursor location.
* src/pgtkterm.c (pgtk_draw_window_cursor): Remove invalid code and
add new code to call the function.
* src/pgtkterm.h: Add the function declaration.
* src/w32fns.c (w32_version_string) [WINDOWSNT]: New function.
* src/w32common.h (w32_version_string) [WINDOWSNT]: Add prototype.
* src/editfns.c (init_editfns) [WINDOWSNT]: Produce a non-nil
string with the OS version.
* src/xdisp.c (produce_stretch_glyph): Truncate the stretch glyph
due to line wrap only when drawing in the text area.
* src/xterm.c (x_draw_stretch_glyph_string):
* src/w32term.c (w32_draw_stretch_glyph_string): Fix the
adjustment of the stretch X and width so that stretch glyphs could
be drawn in the left margin. Reported by Paul W. Rankin
<pwr@bydasein.com>.
The background of the 'child-frame-border' face instead of the
'internal-border' face now controls the color of child frames'
borders.
The 'child-frame-border-width' frame parameter is now used for the
width of child frames' borders instead of internal-border-width',
though we still fall back on using the latter if the former is not
set.
* doc/lispref/frames.texi (Frame Layout): Mention
'child-frame-border' and 'child-frame-border-width'.
(Layout Parameters): Mention 'child-frame-border-width'.
* etc/NEWS: Mention new face 'child-frame-border' and frame
parameter 'child-frame-border-width'.
* lisp/faces.el (child-frame-border): New face.
* src/dispextern.h (enum face_id): Add CHILD_FRAME_BORDER_FACE_ID.
* src/frame.c (Fframe_child_frame_border_width): New function.
(gui_report_frame_params): Add entry for Qchild_frame_border_width.
* src/frame.h (struct frame): New slot child_frame_border_width.
(FRAME_CHILD_FRAME_BORDER_WIDTH): New inlined function.
* src/nsfns.m (ns_set_child_frame_border_width): New function.
(Fx_create_frame): Handle Qchild_frame_border_width parameter.
(ns_frame_parm_handlers): Add ns_set_child_frame_border_width.
* src/nsterm.m (ns_clear_under_internal_border): Handle
CHILD_FRAME_BORDER_FACE_ID.
* src/w32fns.c (w32_clear_under_internal_border): Handle
CHILD_FRAME_BORDER_FACE_ID.
(w32_set_internal_border_width): New function.
(Fx_create_frame): Handle Qchild_frame_border_width parameter.
(w32_frame_parm_handlers): Add w32_set_child_frame_border_width.
* src/xfaces.c (lookup_basic_face, realize_basic_faces): Handle
CHILD_FRAME_BORDER_FACE_ID.
* src/xfns.c (x_set_child_frame_border_width): New function.
(Fx_create_frame): Handle Qchild_frame_border_width parameter.
(x_frame_parm_handlers): Add x_set_child_frame_border_width.
* src/xterm.c (x_clear_under_internal_border)
(x_after_update_window_line): Handle CHILD_FRAME_BORDER_FACE_ID.
While no deadlocks caused by the blocking write end have been reported
yet, marking both ends nonblocking is consistent and also recommended
in the GNU/Linux manpage of 'select'.
* src/process.c (child_signal_init): Mark write end of self-pipe as
nonblocking.
* src/term.c (tty_draw_row_with_mouse_face)
(tty_write_glyphs_with_face): Don't define on MSDOS and WINDOWSNT,
as those have their own implementations of that.
* src/term.c (tty_write_glyphs_with_face): Move definition out of
ifdef block.
* src/xdisp.c (draw_row_with_mouse_face): Now called
unconditionally on all platforms.
* src/image.c: New macro to specify non-scaled mode.
(image_set_transform): Don't skip even if imagemagick.
(imagemagick_load_image): Don't calculate desired size.
