NOTE: *DO NOT* merge this change to the trunk.
* lisp/gnus/gnus-group.el (gnus-read-ephemeral-gmane-group)
(gnus-read-ephemeral-bug-group): Make it work for any number of times
for the case `url-automatic-caching' is set (bug#29008).
This fixes (a follow-up to) bug #28850.
A internal generated form for scanning text to fontify had a LIMIT parameter.
It also locally bound LIMIT to a value possibly beyond the original LIMIT,
allowing point to move beyond the original LIMIT, and to create the wrong side
error. Fix it by checking point is not beyond LIMIT in the outer context
before using it.
* lisp/progmodes/cc-fonts.el (c-make-font-lock-search-form): Add a new
parameter CHECK-POINT which, when non-nil, directs the function to generate a
check on point.
(c-make-font-lock-context-search-function): Invoke the above function with new
argument value t.
Fix intentation problems reported in
https://github.com/mooz/js2-mode/issues/463.
* lisp/progmodes/js.el (js--continued-expression-p):
Check syntax state after /.
(js--multi-line-declaration-indentation):
Check syntax state before "const".
The cause was a scanning over a bracket pair taking us beyond the supplied
LIMIT parameter in c-forward-declarator.
* lisp/progmodes/cc-engine.el (c-forward-declarator): Add three checks (<
(point) limit) whilst dealing with tokens after the declared identifier.
* lisp/progmodes/cc-fonts.el (c-font-lock-declarators): Don't supply a LIMIT
argument to `c-forward-declarator' (twice), since we want to fontify up till
the end of a declarator, not an arbitrary jit-lock chunk end.
* lisp/loadup.el: add `progmodes` to load-path so we can find flymake.el.
* lisp/kmacro.el: Require `replace` since we use query-replace-map.
* lisp/replace.el: Require `text-mode` since we use text-mode-map.
Since 2017-07-25 "ls-lisp: Add an unload function and enable lexical
binding", the non-w32 builds would treat the undeclared
w32-collate-ignore-punctuation variable as lexical.
* lisp/ls-lisp.el (top-level): Declare it as a dynamic variable.
MELPA includes a :commit field in its
packages (https://github.com/melpa/package-build/pull/6). You can use
this to tell if MELPA has processed a recently-merged change. This
commit adds that metadata to the package description buffer.
* lisp/emacs-lisp/package.el: Display commit in package description.
Copyright-paperwork-exempt: yes
* lisp/window.el (window-normalize-buffer): Fix case where
BUFFER-OR-NAME is a string specifying a dead buffer. Fix
doc-string (Bug#28947).
(window-normalize-frame, window-normalize-window): Fix
doc-strings (Bug#28947).
Problem reported by Nelson H. F. Beebe (Bug#28893).
Also see Bug#23748, Bug#9736, and Bug#5735.
* configure.ac (tputs_library): Prefer libcurses to libtermcap,
since OpenIndiana libtermcap lacks tparm.
This approach does mean that keywords that have spaces before them
inside of docstrings aren't filled, but I think this is should be fine
until Bug#28937 is fixed.
* lisp/emacs-lisp/lisp-mode.el (lisp-fill-paragraph): Add a colon to
paragraph-start unconditionally, but require that it follows at least
one space. (Bug#24622)
* test/lisp/emacs-lisp/lisp-tests.el: New tests for Bug#24622 and
Bug#7751.
Use the new function directly in several places where c-forward-token-2
wouldn't move over the last token in the buffer. This caused an infinite loop
in c-restore-<>-properties.
* lisp/progmodes/cc-engine.el (c-forward-over-token-and-ws): New function,
extracted from c-forward-token-2.
(c-forward-token-2): Refactor, calling the new function.
(c-restore-<>-properties): Fix infinite loop.
(c-forward-<>-arglist-recur, c-in-knr-argdecl)
(c-looking-at-or-maybe-in-bracelist): Call the new function directly in place
of c-forward-token-2.
* lisp/progmodes/cc-cmds.el (c-defun-name) Call the new function directly in
place of c-forward-token-2.
* lisp/progmodes/cc-fonts.el (c-font-lock-enclosing-decls): Call the new
function directly in place of c-forward-token-2.
* lisp/xdg.el (xdg-thumb-mtime): Return an Emacs timestamp,
not an integer. This avoids signaling an error on 32-bit
Emacs, where timestamps typically do not fit into fixnums
(Bug#28921).
* lisp/server.el (server-ensure-safe-dir): Put file owner's uid, not
current user's for the wrong owner case. Show expanded file name in
error message.
In 2017-09-13 "Prefer HTTPS to FTP and HTTP in documentation",
gnus-bug-group-download-format-alist was updated to use https for the
debbugs.gnu.org mbox links, but gnus-read-ephemeral-bug-group assumed
http links.
* lisp/gnus/gnus-group.el (gnus-read-ephemeral-bug-group): Use
url-parse functions to get the host name, instead of ad-hoc regexps.
* lisp/progmodes/flymake.el (flymake--mode-line-format): Bind
'mouse-wheel-down-event' and 'mouse-wheel-up-event' instead of
'mouse-4' and 'mouse-5'. Update the tooltip text accordingly, and
remove a stray newline in it.
* lisp/term.el (term-mode, term-char-mode, term-line-mode)
(term-emulate-terminal): Make buffer read-only in 'term-char-mode',
except for the process filter's output. Use 'read-only-mode-hook' to
track and restore the user-set state of 'buffer-read-only' for
'term-line-mode'. (Bug#24837)
(term-char-mode-buffer-read-only): New user option.
(term-line-mode-buffer-read-only): New buffer-local variable.
(term-line-mode-buffer-read-only-update): New function.
(term-char-mode, term-line-mode): Use 'term-set-goto-process-mark'
in pre-command-hook, and 'term-goto-process-mark-maybe' in
post-command-hook to counter-act unexpected changes to point when
using 'term-char-mode'.
(term-char-mode-point-at-process-mark): New user option.
(term-goto-process-mark): New buffer-local variable.
(term-set-goto-process-mark): New function.
(term-goto-process-mark-maybe): New function.
(term-process-mark): New function.
* etc/NEWS: Mention the new behavior and user options.
* lisp/subr.el (make-progress-reporter): Remove no-longer-needed
doc and code for the case where float-time is not supported,
since float-time is always present now.
* lisp/gnus/mail-source.el (mail-source-delete-old-incoming):
The old code had two minor bugs: first, it multiplied by 1/65536
instead of by 1/86400. Second, it called current-time twice and
glued part of the two results together, resulting in a race.
Fix both bugs by simplifying the code and using
time-to-number-of-days.
* src/doc.c (Fdocumentation): Restore the pre-25 behavior
of signalling a void-function error for an undefined function.
This seems cleaner than Emacs 25's "invalid-function: nil" error,
which was (probably) an unintended consequence of changes to
Findirect_function.
