The regex stack limit was being computed as the number of stack entries,
whereas it was being compared with the current size as measured in
bytes. This could cause indefinite looping when nearing the stack limit
if re_max_failures happened not to be a multiple of sizeof
fail_stack_elt_t (Bug #24751).
* src/regex.c (GROW_FAIL_STACK): Compute both current stack size and
limit as numbers of stack entries.
* rst.el (rst-cvs-header, rst-svn-rev, rst-svn-timestamp)
(rst-official-version, rst-official-cvs-rev)
(rst-package-emacs-version-alist): Maintain version numbers.
(rst-forward-indented-block): Fix. Start searching at next
line again. Fixes fontification of comments continuing on the
same line they started.
Emacs now assumes GNU Make, so @SET_MAKE@ is no longer needed.
* Makefile.in, lwlib/Makefile.in, nextstep/Makefile.in:
* src/Makefile.in: Remove @SET_MAKE@.
* lisp/net/tramp-sh.el (tramp-maybe-open-connection):
`start-file-process' shall work when `non-essential' is
non-nil, but there is already an established connection.
<https://github.com/company-mode/company-mode/issues/462>
* lisp/progmodes/sql.el (sql-sqlite-login-params): Allow any name as
SQLite database file name, by default.
(sql-get-login-ext): Fixed read-file-name arguments to provide
path completion even if a database name pattern is customized and to
allow creation of new SQLite database files.
Copyright-paperwork-exempt: yes
* lisp/bookmark.el (bookmark-insert-file-format-version-stamp):
Accept an argument CODING and include a 'coding:' cookie in the
bookmark file preamble.
(bookmark-upgrade-file-format-from-0): Call
'bookmark-insert-file-format-version-stamp' with the file buffer's
encoding, as detected when it was read.
(bookmark-file-coding-system): New variable.
(bookmark-load): Set bookmark-file-coding-system to the encoding
of the loaded file.
(bookmark-write-file): Bind coding-system-for-write to either the
user setting via "C-x RET c" or to the existing file encoding,
defaulting to 'utf-8-emacs'. Update the value of
bookmark-file-coding-system. (Bug#25365)
See Bug#25366.
* src/character.c (blankp): New function for checking Unicode
horizontal whitespace.
* src/regex.c (ISBLANK): Use 'blankp' for non-ASCII horizontal
whitespace.
(BIT_BLANK): New bit for range table.
(re_wctype_to_bit, execute_charset): Use it.
* test/lisp/subr-tests.el (subr-tests--string-match-p--blank): Add
unit test for [:blank:] character class.
* test/src/regex-tests.el (test): Adapt unit test.
* doc/lispref/searching.texi (Char Classes): Document new Unicode
behavior for [:blank:].
* lisp/isearch.el: Add `isearch-scroll' property to
universal-argument-more so that `isearch-allow-scroll' will apply to it
as well.
(isearch-pre-command-hook): Let `isearch-allow-prefix' apply to
`universal-argument-more' as well (Bug#25302).
* autogen.sh (gnulib.mk): Make the Sed script more portable.
* nt/Makefile.in (${srcdir}/gnulib.mk): Adapt the Sed command to
the changes in autogen.sh.
Bash will after each command send ?\032 and the current directory "/tmp"
to inform term.el. Bash output is buffered in 4096 bytes chunks. If a
command outputs roughly 4096 bytes then the end of the first chunk will
be "/tm" (Bug#13350).
* lisp/term.el (term-emulate-terminal): Change the regexp to find the
end of the ?\032 sequence to use \n instead of $, the latter can match
end of string as well.
Copyright-paperwork-exempt: yes
Problem reported by Sam Steingold (Bug#25347).
* autogen.sh: Don't assume 'sed -f-' reads a script from stdin, as
POSIX does not require it and it does not work on Darwin.
* lisp/gnus/nnir.el: Use lexical-binding and cl-lib.
(nnir-retrieve-headers): Use pcase.
(nnir-search-thread): Avoid add-to-list on local variables.
* lisp/gnus/smime.el: Use lexical-binding and cl-lib.
(smime-verify-region): Avoid add-to-list on local variables.
* lisp/mail/undigest.el: Use lexical-binding and cl-lib.
(rmail-digest-parse-mime, rmail-digest-rfc1153)
(rmail-digest-parse-rfc934): Avoid add-to-list on local variable.
* lisp/net/ldap.el (ldap-search): Move init into declaration.
* lisp/net/newst-backend.el (newsticker--cache-add):
Avoid add-to-list on local variables; Simplify code with `assq'.
* lisp/net/zeroconf.el: Use lexical-binding and cl-lib.
(dbus-debug): Remove declaration, unused.
(zeroconf-service-add-hook, zeroconf-service-remove-hook)
(zeroconf-service-browser-handler, zeroconf-publish-service):
Avoid add-to-list and *-hook on local variables.
* lisp/org/org-archive.el (org-all-archive-files):
* lisp/org/org-agenda.el (org-agenda-get-restriction-and-command):
Avoid add-to-list on local variables.
* lisp/org/ox-publish.el (org-publish--run-functions): New function.
(org-publish-projects): Use it to avoid run-hooks on a local variable.
(org-publish-cache-file-needs-publishing): Avoid add-to-list on
local variables.
* lisp/progmodes/ada-prj.el: Use setq instead of (set '...).
(ada-prj-load-from-file): Avoid add-to-list on local variables.
* lisp/progmodes/ada-xref.el (ada-initialize-runtime-library): Simplify.
(ada-gnat-parse-gpr, ada-parse-prj-file-1)
(ada-xref-find-in-modified-ali): Avoid add-to-list on local variables.
* lisp/progmodes/idlw-shell.el (idlwave-shell-update-bp-overlays):
Avoid add-to-list on local variables.
User visible improvements and changes:
* Improve and debug `rst-forward-section` and `rst-backward-section`.
* Auto-enumeration may be used with all styles for list insertion.
* Improve and debug `rst-toc-insert`.
* Adapt change in Emacs to use customization group `text` instead of `wp`.
* Bind `n` and `p` in `rst-toc-mode`.
* `z` in `toc-mode` returns to the previous window configuration.
* Require Emacs version >= 24.1.
Lots of refactorings including:
* Silence byte compiler.
* Use lexical binding.
* Use `cl-lib`.
* Add tests and raise test coverage.
* lisp/emacs-lisp/cl-macs.el (cl-defstruct): Don't signal an error but
emit a warning for those coders who forgot to put a default value in
their slot.
This was proposed by Paul Eggert <eggert@cs.ucla.edu>,
with the purpose of avoiding manual maintenance of
nt/gnulib.mk.
* nt/gnulib-modules-to-delete.cfg: New file.
* nt/Makefile.in (AM_V_GEN, am__v_GEN_, am__v_GEN_0)
(am__v_GEN_1): New variables.
(${srcdir}/gnulib.mk): Rules to generate gnulib.mk from
lib/gnulib.mk and list of modules in gnulib-modules-to-delete.cfg.
* make-dist (nt): Add gnulib-modules-to-delete.cfg to the list of
files to link.
* configure.ac (GNULIB_MK): Compute the value according to $opsys.
* autogen.sh: Create nt/gnulib.mk if it doesn't exist, before
running autoreconf.
* Makefile.in (gnulib_mk): New variable.
($(srcdir)/nt/gnulib.mk): Rule to produce it.
(AUTOMAKE_INPUTS): Use $(gnulib_mk) instead of a literal file
name.
* .gitignore: Add nt/gnulib.mk.
* src/w32.c (acl_errno_valid): Implement it here, as we no longer
build the acl-permissions module from Gnulib.
* lisp/term.el (term-terminal-undecoded-bytes): New variable.
(term-mode): Make it buffer local. Don't make `term-terminal-parameter'
buffer-local twice.
(term-emulate-terminal): Check for bytes of incompletely decoded
characters, and save them until the next call when they can be fully
decoded (Bug#25288).
* test/lisp/filenotify-tests.el (file-notify--test-monitors):
New variable.
(file-notify--test-cleanup, file-notify--test-monitor): Use it.
(file-notify--test-read-event, file-notify-test02-events)
(file-notify-test04-file-validity): Handle "gvfs-monitor-dir.exe".
(file-notify-test03-autorevert)
(file-notify-test08-watched-file-in-watched-dir):
Set `file-notify--test-desc' for proper work of
`file-notify--test-monitor'. (Bug#21804)
* lisp/textmodes/reftex.el (reftex-compile-variables): Change the
section regexp so that it accepts lines starting with the comment
character. (tiny change)
* lisp/textmodes/reftex-parse.el (reftex-parse-from-file): Filter
gathered toc entries, accepting a commented entry if and only if the
source file is a ".dtx" file. (tiny change)
The current tar-mode doesn't really support unpacking symlinks, it
simply creates an empty file of the same name.
* lisp/tar-mode.el (tar--describe-as-link): New function extracted from
`tar--check-descriptor'.
(tar-untar-buffer): Use it to warn about imperfectly untarred link
files.
* lisp/progmodes/sh-script.el (sh-mode): Remove local setting of
`skeleton-end-hook', `skeleton-insert' already does `newline-and-indent'
and also respects `skeleton-end-newline' (Bug#16634).
Since tmux version 2.1, new tmux terminfos are shipped due to oddities
with xterm and screen terminfos. This is simply a duplication of
term/screen.el with screen -> tmux.
* lisp/term/tmux.el: New file.
JSON requires that such pairs be treated as UTF-16 surrogate pairs, not
individual code points; cf. Bug #24784.
* lisp/json.el (json-read-escaped-char): Fix decoding of surrogate
pairs.
(json--decode-utf-16-surrogates): New defun.
* test/lisp/json-tests.el (test-json-read-string): Add test for
surrogate pairs.
* admin/merge-gnulib: Remove m4/wint_t.m4 when merging.
Fix typo so that warn-on-use.m4 is removed too.
* configure.ac (gt_TYPE_WINT_T): New macro, replacing Gnulib’s.
* m4/wint_t.m4: Remove.
This incorporates:
2016-12-19 stdint: Fix WINT_MAX to match wint_t on mingw
2016-12-18 getopt: Fix link error for users of getopt() in <unistd.h>
2016-12-17 getlogin: Port to newer mingw
2016-12-17 stdint: Fix WINT_MAX to match wint_t on MSVC
2016-12-17 Avoid redefinition errors on MSVC
* lib/getopt.in.h, lib/stdint.in.h, lib/stdio.in.h, lib/unistd.in.h:
* m4/stdint.m4, m4/unistd_h.m4:
Copy from gnulib.
* lib/gnulib.mk, m4/gnulib-comp.m4: Regenerate.
Plus, this commit updates the indenting on copyright notices to
match that of gnulib.
