This is easier to read and maintain, and makes the state explicit.
It is a pure refactoring; the compiled code should be equivalent.
* src/print.c (PRINTPREPARE, PRINTFINISH): Replace with...
(struct print_context, print_prepare, print_finish): ...these new
functions and explicit state in a struct.
(Fwrite_char, write_string, Fterpri, Fprin1, Fprin1_to_string)
(Fprinc, Fprint): Adapt callers.
* lisp/obsolete/tpu-edt.el: Remove autoloads of obsolete library
tpu-extras.el.
* lisp/obsolete/tpu-extras.el: Don't set generated-autoload-file
to tpu-edt.el.
* lisp/filesets.el:
* lisp/mh-e/mh-*.el:
* lisp/progmodes/python.el (python): Don't set options that are
already set in .dir-locals.el.
* lisp/textmodes/ispell.el: Remove redundant local variables.
* lisp/gnus/nnmaildir.el: Remove local variables that conflicts with
Emacs defaults.
* src/xfns.c (x_relative_mouse_position)
(Fx_mouse_absolute_pixel_position, compute_tip_xy): Use
x_query_pointer instead of XQueryPointer.
* src/xterm.c (xi_populate_device_from_info): Set new attachment
field.
(xi_handle_focus_change): Set client pointer.
(xi_focus_handle_for_device, xi_handle_interaction): Use
attached keyboard device to handle focus.
(x_query_pointer): New function.
(XTmouse_position, x_scroll_bar_report_motion)
(x_horizontal_scroll_bar_report_motion, handle_one_xevent): Use
x_query_pointer instead of XQueryPointer.
(x_term_init): Initialize client pointer device.
* src/xterm.h (struct xi_device_t): New field `attachment'.
(struct x_display_info): New field `client_pointer_device'.
* src/xterm.c (xi_handle_focus_change): Pacify a wrong warning on
gcc (Debian 10.2.1-6).
xterm.c: In function ‘xi_handle_focus_change’:
xterm.c:12564:25: warning: ‘source’ may be used uninitialized in
this function [-Wmaybe-uninitialized]
12564 | ie.device = source->name;
| ~~~~~~^~~~~~
* src/xdisp.c (move_it_vertically_backward): Fix the case of long
bidi-reordered lines when starting from its last (top-most) screen
line.
(move_it_by_lines): Fix logic of detection when going by lines
overshoots.
* lisp/vc/ediff-init.el (ediff-user-grabbed-mouse): Declare unused
function obsolete. As far as I can tell, this has been unused
since 1996.
* lisp/vc/ediff-wind.el (ediff-mouse-pixel-position): Make
variable obsolete. It is only used in the above obsolete funct
* lisp/vc/ediff-mult.el (ediff-filegroup-action)
(ediff-registry-action):
* lisp/vc/ediff-util.el (ediff-setup): Don't set above unused
variable.
* lisp/vc/ediff-init.el (ediff-spy-after-mouse): Make obsolete.
* lisp/vc/ediff-util.el (ediff-setup): Don't add above obsolete
function to pre-command-hook.
* test/lisp/eshell/em-pred-tests.el
(eshell-parse-file-name-attributes): Use t instead of nil as
second argument to 'time-convert'.
Suggested by Stefan Monnier <monnier@iro.umontreal.ca>.
* lisp/url/url-util.el (url-display-message): New function.
(url-display-percentage): Make obsolete in favor of
url-display-message.
* lisp/url/url-http.el
(url-http-content-length-after-change-function):
Prefer 'url-display-message' to 'url-display-percentage'.
(url-http-content-length-after-change-function)
(url-http-chunked-encoding-after-change-function):
Remove ineffectual calls to 'url-display-percentage'.
7ffcba4213 * etc/PROBLEMS: Problems with Alacritty and Emoji. (Bug#5...
df95118e83 Link from (emacs)Init Syntax to (elisp)Introduction
7793cf0c3d Don't mention removed XEmacs support in reftex manual
3d1cf1b47a Don't mention removed XEmacs support in idlwave manual
7ff60c62a9 ; * admin/make-tarball.txt: Fix last change.
2fd5590d0c ; * lisp/play/fortune.el (fortune-ask-file): Doc fix.
'User variables' were made obsolete in Emacs 24 along with
user-variable-p; the sign of the position in (#$ . POS) hasn't
mattered since.
* lisp/emacs-lisp/bytecomp.el (byte-compile-output-docform):
Don't emit negative position when doc string starts with `*`.
* src/lread.c (get_lazy_string): Explain.
When load-force-doc-strings is true, read (#$ . POS) as the (unibyte)
string referred to. This feature was lost by mistake in the recent
nonrecursive reader rewrite.
Noticed by Stefan Monnier.
* src/lread.c (get_lazy_string): New function (code mostly recycled
from an old version).
(read0): Detect (#$ . FIXNUM) and retrieve the string if appropriate.
* test/src/lread-resources/lazydoc.el:
* test/src/lread-tests.el (lread-force-load-doc-strings):
New test.
Fix two problems with our handling of X timestamps
1) We're not properly updating the X interaction timestamp after
receiving certain input events, and
2) X events sent in response to emacsclient commands get stale
timestamps because the timestamp tracking doesn't take into account
that interactions with the user can occur outside the X input
event channel.
* src/xterm.c:
(x_display_set_last_user_time_1): New function.
(x_display_set_last_user_time): Call it.
(x_ewmh_activate_frame): Refactor.
(x_focus_frame): Don't call XSetInputFocus if we can use EWMH activation.
(server_timestamp_predicate): New function.
(x_get_server_time): New function.
(x_note_oob_interaction): New function.
(x_create_terminal): Register new function as terminal hook.
* src/termhooks.h: New hook: note_oob_interaction_hook.
* src/gtkutil.h:
(xg_set_user_timestamp): Declare.
* src/gtkutil.c:
(xg_set_user_timestamp): New function.
* src/frame.c:
(Fframe_note_oob_interaction): New function.
(syms_of_frame): Register it.
* lisp/server.el:
(server-switch-buffer): Call frame-note-oob-interaction when user
requests frame be raised.
Also replace some regexp searches which had caused regexp engine stack
overflows with simple end-of-line calls.
* lisp/progmodes/cc-fonts.el (c-make-syntactic-matcher)
(c-make-font-lock-search-form): Start the generated functions with a skipping
of comments and strings.
(c-make-font-lock-BO-decl-search-function): Start the generated function with
a (fast) movement to the start of any literal.
* lisp/progmodes/cc-mode.el (c-before-change-check-unbalanced-strings)
(c-after-change-mark-abnormal-strings): Replace complicated regexp searches
for end of logical line with basic Lisp functions.
This used to be for XEmacs and old version compatibility, but nowadays
everything in this file is obsolete.
* lisp/mh-e/mh-compat.el: Move from here...
* lisp/obsolete/mh-compat.el: ...to here.
* lisp/allout.el (allout-outlinify-sticky): Rename from
'outlineify-sticky'. Make old name into obsolete alias.
(outlinify-sticky): Make into obsolete function alias for
'allout-outlinify-sticky'.
(allout-mode): Doc fix.
Ubuntu 16.04 LTS reached end of life on April 30th, 2021.
Furthermore, test failures don't seem significant enough to warrant an
entry. It's rather cosmetic.
* etc/PROBLEMS: Delete entry on test failures on Ubuntu 16.04.
Eventually we want to be able to generate "transient.texi" from
"transient.org", without having to either give up on idiomatic texinfo
or making it much more painful to maintain the org file.
We are much closer to that now, but there are still a few areas where
additional work is needed. This was mostly accomplished by using Org
macros.
The most significant outstanding issue is that the generated
references don't yet look like an experienced texinfo author like Eli
would like them to look. Additionally it is not yet possible to use a
macro that produces @dots{} in the places Eli added them, and in Org
code blocks it is not possible to use macros, so we cannot have
@var{...} appear in "@lisp ... @end lisp". The last issue probably
cannot be changed on Org's side, but since there are only two such
code blocks, this might be a situation where the compromise has to
come from the texinfo side. There are also three other very minor
and inconsequential differences.
For now I have regenerated the texinfo file from the org file and then
discarded the differences mentioned in the previous paragraph.
The process of merging (1) Eli's changes to the texinfo file
(including, but certainly not limited to markup), (2) changes to the
org source (updated content, formatting changes backported earlier,
fixes for formatting changes Eli did not fix, etc.) and (3) changes to
the code that converts the org source to texinfo, was very laborious
and painful. In essence, this amounted to a (at least) three-way
merge across three different languages and three repositories.
I tried very hard to not waste any of the effort Eli had put into
fixing up the generated texinfo file. I.e., I went back and forth
making improvements to the org source, implementing org macros,
regenerating the texinfo and comparing the remaining difference, and
creating commits on both sides. This resulted in a dozen commits on
both sides and took me well over a day. I could have put in even more
effort to absolutely ensure nothing at all is lost in the process, but
I think that would have amounted to a colossal waste of my time.
Going forward, if you find unidiomatic texinfo, then please don't fix
each instance. Instead write me an email, explaining what the problem
is. You are welcome to make limited fixes to the content or fix
one-of markup issue in the texinfo file; those are relatively simple
to backport in comparison.
