Paul Eggert [Mon, 30 May 2016 23:09:25 +0000 (16:09 -0700)]
Omit IF_LINT code that no longer seems needed
Nowadays GCC is smarter, or the Emacs code has mutated, or both,
and now is as good a time as any to remove uses of IF_LINT that
now seem to be unnecessary.
* lib-src/emacsclient.c (set_local_socket):
* lib-src/movemail.c (main) [MAIL_USE_MAILLOCK && HAVE_TOUCHLOCK]:
* src/buffer.c (fix_start_end_in_overlays, fix_overlays_before):
* src/casefiddle.c (casify_region):
* src/charset.c (load_charset_map):
* src/coding.c (decode_coding_object, encode_coding_object):
* src/data.c (Fmake_variable_buffer_local, Fmake_local_variable)
(cons_to_unsigned, cons_to_signed):
* src/frame.c (make_frame, x_set_frame_parameters):
* src/keyboard.c (read_event_from_main_queue):
* src/regex.c (regex_compile):
* src/syntax.c (back_comment):
* src/window.c (Frecenter):
* src/xfaces.c (Fx_list_fonts):
Remove IF_LINT that no longer seems necessary.
* src/image.c (png_load_body, jpeg_load_body): Simplify use of IF_LINT.
* src/keyboard.c (read_char): Use IF_LINT (volatile) rather than
a pragma dance to pacify GCC -Wclobbered.
* src/xdisp.c (x_produce_glyphs): Rewrite to avoid need for IF_LINT.
* src/xterm.c (x_connection_closed): Now _Noreturn, which should
mean we do not need IF_LINT any more.
(x_io_error_quitter): Now _Noreturn. Put an 'assume (false)’
at the end, to forestall warnings from older compilers.
Paul Eggert [Mon, 30 May 2016 23:09:25 +0000 (16:09 -0700)]
Port --enable-gcc-warnings to Cygwin, FreeBSD
These platforms have a bug where _Noreturn is empty when 'lint' is
defined. Problem reported by Ken Brown (Bug#23640).
* configure.ac (GCC_LINT): Rename from 'lint'.
* src/conf_post.h (IF_LINT): Use GCC_LINT, not just 'lint’.
Paul Eggert [Mon, 30 May 2016 23:09:25 +0000 (16:09 -0700)]
Update from gnulib
This incorporates:
2016-05-30 Use GCC_LINT, not lint
2016-05-29 secure_getenv: Port to many more platforms.
* doc/misc/texinfo.tex, lib/secure_getenv.c, lib/verify.h:
* m4/secure_getenv.m4: Copy from gnulib.
Allow preserving EXIF rotations when sending HTML messages
* lisp/gnus/mml.el (mml--possibly-alter-image): Allow image
rotation if you have exiftool installed and the image format
supports it.
(mml-expand-html-into-multipart-related): Use it.
(mml-buffer-substring-no-properties-except-some): Renamed and
copy display properties, too.
* lisp/gnus/message.el (message-toggle-image-thumbnails): Use
`insert-image' instead of `put-image' to make it possible to
edit the resulting text in a sensible manner.
Alan Mackenzie [Sun, 29 May 2016 12:22:17 +0000 (12:22 +0000)]
Rationalize the use of c-new-BEG and c-new-END in CC Mode.
Remove the now redundant c-old-BOM and c-old-EOM.
* lisp/progmodes/cc-engine.el (c-macro-cache-syntactic): Change and simplify
meaning.
(c-macro-cache-no-comment): New variable.
(c-invalidate-macro-cache, c-beginning-of-macro, c-end-of-macro): incorporate
the new c-macro-cache-no-comment.
(c-syntactic-end-of-macro): Make better use of c-macro-cache-syntactic.
(c-no-comment-end-of-macro): New function.
* lisp/progmodes/cc-langs.el (c-before-font-lock-functions): Add
c-extend-font-lock-region-for-macros to C/C++/ObjC value.
* lisp/progmodes/cc-mode.el (c-old-BOM, c-old-EOM): Remove.
(c-extend-region-for-CPP): Put results in c-new-BEG/END rather than
c-old-BOM/EOM.
(c-extend-font-lock-region-for-macros): Simplify meaning, no longer returning
a cons for the new region, since the function is now called as an after-change
function. No longer adjust c-new-END for the length of inserted/deleted
text. Move the size restrictions on macros to here from
c-neutralize-syntax-in-and-mark-CPP.
(c-neutralize-syntax-in-and-mark-CPP): No longer adjust c-new-BEG/END here.
Use c-no-comment-end-of-macro rather than c-syntactic-end-of-macro to find the
upper boundary to "neutralize" syntactically obtrusive characters.
(c-change-expand-fl-region): Don't set c-new-END to next BOL when already at
one.
Eli Zaretskii [Sat, 28 May 2016 15:33:50 +0000 (18:33 +0300)]
Fix 'next-line' and 'previous-line' during keyboard macro
* lisp/simple.el (line-move-visual): Handle the case when point
moves outside of the visible portion of the buffer while executing
a keyboard macro. (Bug#13452) (Bug#23551) (Bug#23555)
Michael Albinus [Sat, 28 May 2016 10:14:58 +0000 (12:14 +0200)]
Sync with Tramp repository.
* lisp/net/tramp-adb.el (tramp-adb-handle-start-file-process):
* lisp/net/tramp-sh.el (tramp-sh-handle-start-file-process)
* lisp/net/tramp-smb.el (tramp-smb-handle-start-file-process):
Create BUFFER if it doesn't exist. Reported by Nikolay
Kudryavtsev <nikolay.kudryavtsev@gmail.com>.
* lisp/net/tramp-gvfs.el (tramp-gvfs-file-attributes)
(tramp-gvfs-file-attributes-with-gvfs-ls-regexp)
(tramp-gvfs-file-attributes-with-gvfs-info-regexp): New defconst.
(tramp-gvfs-get-directory-attributes)
(tramp-gvfs-get-root-attributes)
(tramp-gvfs-get-file-attributes): New defun.
(tramp-gvfs-handle-file-attributes)
(tramp-gvfs-handle-file-name-all-completions): Use them.
(tramp-gvfs-handle-file-directory-p): Use `file-truename'.
(tramp-gvfs-maybe-open-connection): Set :noquery flag for process.
(tramp-gvfs-send-command): Add locale settings to `process-environment'.
(top): Suppress D-Bus error messages during zeroconf initialization.
* lisp/net/tramp.el (tramp-unknown-id-string)
(tramp-unknown-id-integer): New defconst.
(tramp-check-cached-permissions)
* lisp/net/tramp-gvfs.el (tramp-gvfs-handle-file-attributes)
* lisp/net/tramp-sh.el (tramp-do-file-attributes-with-ls)
(tramp-convert-file-attributes, tramp-get-remote-uid)
(tramp-get-remote-gid): Use them.
* lisp/net/tramp-sh.el (tramp-display-escape-sequence-regexp):
Rename from `tramp-color-escape-sequence-regexp'.
(tramp-sh-handle-insert-directory)
(tramp-convert-file-attributes): Adapt callees.
(tramp-device-escape-sequence-regexp): New defconst.
(tramp-wait-for-output): Use it. Reported by Matthías Páll
Gissurarson <icetritlo@gmail.com>.
Paul Eggert [Sat, 28 May 2016 01:16:24 +0000 (18:16 -0700)]
Don’t document declare-function internals
Suggested by Stefan Monnier in:
http://lists.gnu.org/archive/html/emacs-devel/2016-05/msg00618.html
* doc/lispref/functions.texi (Declaring Functions):
* lisp/subr.el (declare-function):
* lisp/emacs-lisp/bytecomp.el:
(byte-compile-macroexpand-declare-function):
Document as (fn file &optional arglist fileonly)
even though it is really (fn file &rest args).
Paul Eggert [Fri, 27 May 2016 20:39:34 +0000 (13:39 -0700)]
Robustify stack-size calculation
* src/emacs.c: Include getpagesize.h.
(main): Check for integer overflow when computing stack size.
Round new rlim_cur to pagesize boundary on all platforms, as this
is easy and would have prevented Bug#23622. If setrlimit
fails, use current limit to determine re_max_failures.
Ken Brown [Fri, 27 May 2016 18:55:34 +0000 (14:55 -0400)]
Adjust filenotify-tests on Cygwin
* test/lisp/filenotify-tests.el (file-notify--test-timeout):
Remove special case for Cygwin.
(file-notify-test07-backup): Update expected results on Cygwin.
Mark Oteiza [Fri, 27 May 2016 01:47:18 +0000 (21:47 -0400)]
Preserve buffer point in windows by default (Bug#4041).
* doc/lispref/windows.texi: Mention new default.
* etc/NEWS: Mention new default.
* lisp/window.el (switch-to-buffer-preserve-window-point): Default to t.
Paul Eggert [Fri, 27 May 2016 02:10:26 +0000 (19:10 -0700)]
Fix byte-compiler pacification for declare-function
Problem reported by Michael Heerdegen in:
http://lists.gnu.org/archive/html/emacs-devel/2016-05/msg00590.html
* lisp/emacs-lisp/bytecomp.el:
(byte-compile-macroexpand-declare-function):
Revert signature to previous value.
* lisp/subr.el (declare-function): Change signature to
match the reverted signature used in the byte compiler.
Paul Eggert [Thu, 26 May 2016 20:02:07 +0000 (13:02 -0700)]
Repair damage from emacs-25 merge
* test/lisp/character-fold-tests.el:
Remove, as it was renamed to test/lisp/char-fold-tests.el.
* test/lisp/emulation/viper-tests.el:
Rename from test/automated/viper-tests.el.
Michael Albinus [Thu, 26 May 2016 17:57:13 +0000 (19:57 +0200)]
Fix Bug#23614.
* lisp/net/tramp-sh.el (tramp-maybe-open-connection):
Let-bind `process-coding-system-alist' in order to suppress
the value for "cmdproxy".
* lisp/net/tramp.el (tramp-encoding-shell):
`w32-shell-name' is a function.
(tramp-encoding-command-switch)
(tramp-encoding-command-interactive): Use `w32-shell-dos-semantics'.
* test/lisp/net/tramp-tests.el (tramp-test24-file-name-completion):
Suppress some tests on MS Windows. (Bug#23614)
Stephen Berman [Thu, 26 May 2016 09:30:29 +0000 (11:30 +0200)]
hl-line.el: Fix flickering of highlighted line (bug#23510)
* lisp/hl-line.el (hl-line-maybe-unhighlight)
(global-hl-line-highlight-all)
(global-hl-line-maybe-unhighlight): New functions.
(hl-line-overlay-buffer): New variable.
(hl-line-mode): Use it. Replace hl-line-unhighlight on
pre-command-hook by hl-line-maybe-unhighlight on
post-command-hook, to prevent hl-line from flickering.
Adjust document string.
(global-hl-line-mode): Use global-hl-line-highlight-all to
simultaneously highlight the current line in all live windows.
Replace global-hl-line-unhighlight on pre-command-hook by
global-hl-line-maybe-unhighlight on post-command-hook, to
prevent global-hl-line from flickering. Remove
global-hl-line-unhighlight from change-major-mode-hook on
disabling the mode. Adjust document string.
Vincent Belaïche [Thu, 26 May 2016 09:03:21 +0000 (11:03 +0200)]
Correct old cell name unbinding when renaming cell.
Bug is to unbind old cell names when renaming a cell with
'makunbound'. when the old cell name is of A1 type, then
'kill-local-variable' must be used instead, so that only the current
spreadsheet is affected. When the old cell name is a renamed cell,
then 'ses--unbind-cell-name' must be used in order to remove the old
name from the name hashmap.
* ses.el (ses-rename-cell): check consistency of cell symbol from
text-property and from array object. Instead of 'makunbound', use
either 'ses--unbind-cell-name' or 'kill-local-variable' depending on
whether the cell old name is a named cell or an A1 type cell
Stefan Monnier [Thu, 26 May 2016 02:58:18 +0000 (22:58 -0400)]
Make autoloads populate a new definition-prefixes table
* lisp/subr.el (definition-prefixes): New hash table.
(register-definition-prefixes): New function.
* lisp/emacs-lisp/autoload.el (autoload-compute-prefixes): New var.
(autoload--split-prefixes-1, autoload--split-prefixes)
(autoload--make-defs-autoload): New functions.
(autoload-defs-autoload-max-size, autoload-popular-prefixes): New vars.
(autoload-generate-file-autoloads): Obey autoload-compute-prefixes.
(update-directory-autoloads): Don't touch loaddefs.el if the set of
autoloads hasn't changed (i.e. if only the timestamp would change).
* lisp/loadup.el: Purify definition-prefixes.
* lisp/w32-fns.el: Keep name space clean.
(w32-set-default-process-coding-system): Rename from
set-default-process-coding-system.
(w32-set-system-coding-system): Rename from set-w32-system-coding-system.
Eli Zaretskii [Tue, 24 May 2016 16:04:16 +0000 (19:04 +0300)]
Avoid aborting due to errors in arguments of 'set-face-attribute'
* src/xfaces.c (Finternal_set_lisp_face_attribute): Check the
FRAME argument before using it. This avoids gratuitous aborts in
Emacs built with --enable-checking when the luser was unlucky
enough to get the argument list in wrong order.
Stefan Monnier [Tue, 24 May 2016 00:09:33 +0000 (20:09 -0400)]
* lisp/url/url-http.el: Use lexical-binding
(url-http-simple-after-change-function): Use buffer-size rather than `nd'.
(url-http-wait-for-headers-change-function): Remove unused var
`content-length'.
(url-http): Remove unused vars `host' and `port'.
Stefan Monnier [Mon, 23 May 2016 19:12:24 +0000 (15:12 -0400)]
* lisp/image.el: Use lexical-binding
(image-scaling-factor, imagemagick-types-inhibit)
(imagemagick-enabled-types): Remove redundant :group.
(image--get-image): Apply de-Morgan and use car-safe.
(image-compute-scaling-factor): Use the argument.
; 1) put-image does not add a display property to its overlay
; 2) it does add a display text property
; 3) an overlay is not a cons whose car is 'image
Paul Eggert [Mon, 23 May 2016 15:56:42 +0000 (08:56 -0700)]
Don’t use only last protocol from getaddrinfo
Problem reported by Ken Brown in:
http://lists.gnu.org/archive/html/emacs-devel/2016-05/msg00483.html
* src/process.c (conv_addrinfo_to_lisp): New function.
(connect_network_socket): Arg is now a list of addrinfos, not
merely IP addresses. All uses changed. Use protocol from
each addrinfo.
(Fmake_network_process): Accumulate protocols into addrinfos
rather than just using the last one found.
(check_for_dns): Accumulate protocols here, too.
* src/process.h (struct Lisp_Process): Remove ai_protocol;
no longer needed.
Paul Eggert [Mon, 23 May 2016 15:13:03 +0000 (08:13 -0700)]
Fix seq requirement that broke bootstrap
* lisp/image.el (image--get-image): Require seq here, not at the
top level, to avoid ‘(require seq) while preparing to dump’ while
bootstrapping. Suggested by Tino Calancha in:
http://lists.gnu.org/archive/html/emacs-devel/2016-05/msg00477.html
Phillip Lord [Sat, 21 May 2016 09:06:12 +0000 (10:06 +0100)]
Fix viper undo breakage from undo-boundary changes
* lisp/simple.el (undo-auto-disable-boundaries): New variable
* lisp/emulation/viper-init.el (viper-undo-in-complex-command,
viper-undo-needs-adjustment): Rename
variable to reflect new purpose,
(viper-buffer-undo-list-mark): Remove
* lisp/emulation/viper-cmd.el (viper-set-complex-command-for-undo):
Add undo using `undo-boundary', disable default undo-boundary
addition.
* lisp/emulation/viper-cmd.el
(viper-complete-complex-command-for-undo,viper-adjust-undo):
Change function name, remove old undo list manipulation.
* lisp/emulation/viper-cmd.el (viper-undo,viper-repeat,
viper-change-state-to-vi,viper-change-state):
Update for function name change.
Kaushal Modi [Mon, 23 May 2016 00:48:33 +0000 (17:48 -0700)]
Fix reference to obsolete fn ps-eval-switch
lpr-eval-switch is functionally the same as ps-eval-switch,
which was obsoleted in 2cdeb903c57126d3ad5f0cbd72e182584b76ee29.
So use that instead (Bug#19717).
* lisp/printing.el (pr-switches): Use lpr-eval-switch instead of
ps-eval-switch.
Paul Eggert [Mon, 23 May 2016 00:29:58 +0000 (17:29 -0700)]
Do not trash symlinks to init file
If the user’s init file is a symbolic link, do not break the link
when initializing the package system. Problem reported by Jackson
Hamilton (Bug#23050).
* lisp/emacs-lisp/package.el (package--ensure-init-file):
Bind find-file-visit-truename when visiting the init file, and
save and restore the buffer name the way cus-edit does in a
similar situation (Bug#454).
Paul Eggert [Sun, 22 May 2016 20:38:53 +0000 (13:38 -0700)]
Bring back xterm pasting with middle mouse
Problem reported by Jan Synáček.
Solution suggested by Stefan Monnier (Bug#23519).
* lisp/isearch.el (isearch-mode-map): Add a binding for xterm-paste.
(xterm--pasted-text): New decl.
(isearch-xterm-paste): New function.
* lisp/term/xterm.el (xterm--pasted-text): New function,
taken from xterm-paste internals.
(xterm-paste): Use it.
Eli Zaretskii [Sun, 22 May 2016 16:20:28 +0000 (19:20 +0300)]
Provide workaround for xftfont rendering problem
* src/xftfont.c (syms_of_xftfont) <xft-font-ascent-descent-override>:
New variable.
(xftfont_open): Use it to work around problems with rendering some
fonts. (Bug#23360)
Paul Eggert [Sun, 22 May 2016 00:04:44 +0000 (17:04 -0700)]
Prefer SOCK_NONBLOCK to O_NONBLOCK
* src/process.c (SOCK_NONBLOCK): Define to 0 if not already defined.
(connect_network_socket): Create the socket with SOCK_NONBLOCK, to
avoid an fcntl with O_NONBLOCK if SOCK_NONBLOCK works. Put the
SOCK_DGRAM check a bit later, to keep the logic cleaner, as
the order does not matter here.
projects / emacs.git / log