* lisp/emacs-lisp/ert.el (ert-load-file-name): New defvar.
(ert-write-junit-test-report): Use it. Make detection of selector
more robust. Protect calls of `xml-escape-string' with 'noerror.
* test/infra/Makefile.in (subdir_template): Use "make -k ...".
* test/infra/gitlab-ci.yml (test-filenotify-gio):
Adapt make_params in order to get a JUnit test report.
(test-native-comp-speed0): Use "make -k ...".
* test/infra/test-jobs.yml: Regenerate.
* test/lisp/progmodes/perl-mode-tests.el (top):
Set `ert-load-file-name'.
* src/xterm.c (x_unlink_touch_point): Return if touchpoint was
actually unlinked.
(handle_one_xevent): Catch and ignore errors during touch
sequence grabbing.
* src/xdisp.c (window_text_pixel_size): Fix whitespace and braces.
* src/xdisp.c (Fwindow_text_pixel_size):
* etc/NEWS:
* doc/lispref/display.texi (Size of Displayed Text): Improve
wording of the documentation of the new argument of
'window-text-pixel-size'.
* doc/lispref/display.texi (Size of Displayed Text): Update
documentation.
* etc/NEWS: Announce new argument.
* src/xdisp.c (window_text_pixel_size): Allow controlling if
the iterator's ascent and descent will be appended to the
pixel height returned. All callers changed.
(Fwindow_text_pixel_size): New argument `ignore-line-at-end'.
All callers changed.
* src/xdisp.c (handle_display_prop): Pass the window to
'get_char_property_and_overlay', not the buffer. (Bug#52385)
The assignment of the buffer to OBJECT was moved to before
the call to 'get_char_property_and_overlay', for unknown
reasons, as part of installing the support for the
'min-width' space spec.
Emacs might get an untracked TouchUpdate event in the unlikely
event of a device hierarchy change during a touch sequence, in
which case all devices and touchpoints will be reset.
* src/xterm.c (handle_one_xevent): Don't abort when receiving
a TouchUpdate that isn't part of a touch sequence.
Avoid the use of the old (eql VAL) specializer, preferring the new
(eql EXP) form.
* lisp/emacs-lisp/multisession.el (multisession-backend-value)
multisession--backend-set-value, multisession--backend-values)
multisession--backend-delete, multisession-backend-value)
multisession--backend-set-value, multisession--backend-values)
multisession--backend-delete): Quote the symbol passed to `eql` specializer.
* lisp/net/tramp-adb.el (tramp-adb-ls-output-time-less-p):
Revert previous change since Tramp code should still be backward
compatible to Emacs 26. But add a comment about this.
Problem reported by Michael Albinus in:
https://lists.gnu.org/r/emacs-devel/2021-12/msg01623.html
* src/font.c (font_delete_unmatched): The kludge of allowing
inexact matches of the font weight is now used for non-NTGUI
platforms as well. (Bug#52493)
* lisp/emacs-lisp/multisession.el: Don't require url -- this pulls
in gnus-util, which requires rmail, which depends on
rmail-loaddefs being generated, which our Makefiles doesn't guarantee.
* lisp/epa-ks.el (epa-ks--parse-buffer): Omit unnecessary calls to
seconds-to-time, and prefer the unaliased name time-convert when
calls are necessary.
Do not warn about timestamps like (1 . 1000). This warning was added
in Emacs 27 as a temporary transition aid, and has now served its
purpose. These timestamps, which Emacs 26 and earlier treated as (HI
. LO) instead of as (TICKS . HZ), were never generated by Emacs
primitives, and in practice the warning seems to have been triggered
only by test cases designed to generate it.
* src/timefns.c (WARN_OBSOLETE_TIMESTAMPS): Remove.
All uses changed to assume it’s false.
(decode_lisp_time): Simplify by taking a bool instead of an
integer bitmask. All uses changed.
* lisp/calendar/icalendar.el (icalendar--decode-isodatetime):
Use pcase instead of cl-case, as cl-macs might not be loaded.
Otherwise, a build complains “calendar/icalendar.el:618:57:
Warning: ‘90’ is a malformed function” and some
icalendar--decode-isodatetime tests fail. This fixes a bug
introduced in 2021-11-21T08:18:57Z!dick.r.chiang@gmail.com.
