João Távora [Mon, 24 Dec 2018 12:36:10 +0000 (12:36 +0000)]
Correctly implement two electric.el tests
* test/lisp/electric-tests.el (electric-pair-mode-newline-between-parens)
(electric-layout-mode-newline-between-parens-without-e-p-m):
Correctly code these two tests.
João Távora [Mon, 24 Dec 2018 12:15:46 +0000 (12:15 +0000)]
Clean up description of new semantics of electric-layout-rules
* lisp/electric.el (electric-layout-rules): Describe function
elements.
(electric-layout-post-self-insert-function-1): Fix handling of
function elements. Describe that e-l-m is good enough to handle
electric-pair-open-newline-between-pairs.
João Távora [Sat, 22 Dec 2018 22:27:27 +0000 (22:27 +0000)]
Test electric.el in plainer c-mode without c-electric-{brace,paren}
If cc-mode ever drops/changes its usage of c-electric-{brace,paren}
test that electric pairing, layout and indentation based on
electric.el keep working.
* test/lisp/electric-tests.el
(electric-layout-int-main-kernel-style)
(electric-layout-int-main-allman-style): Change order of after and
after-stay.
(plainer-c-mode): New test mode.
(electric-modes-in-c-mode-with-self-insert-command): New test.
João Távora [Sat, 22 Dec 2018 21:05:23 +0000 (21:05 +0000)]
Handle entries of multiple symbols in electric-layout-rules
Instead of allowing multiple rules in electric-layout-rules, allow
rules to specify multiple symbols. This should be entirely
backward-compatible to existing customizations of that variable.
* lisp/electric.el (electric-layout-rules): a single entry can
specify multiple symbols.
(electric-layout-post-self-insert-function-1): rework.
* test/lisp/electric-tests.el: Update tests to work with new semantics
of electric-layout-rules.
This should now fix more problems reported in (bug#33794) regarding
insertion of newlines before and after the opening brace. Write two
automated tests.
Also provide a new electric-layout-local-mode for testing.
* lisp/electric.el (electric-layout-post-self-insert-function-1):
New function that does the work for
electric-layout-post-self-insert-function-1.
(electric-layout-local-mode): New minor mode.
* test/lisp/electric-tests.el (electric-layout-int-main-kernel-style)
(electric-layout-int-main-allman-style): Add two tests.
João Távora [Fri, 21 Dec 2018 18:00:08 +0000 (18:00 +0000)]
Extend electric-layout-mode to handle more complex layouts
Also, have it play nice with electric-pair-mode.
Multiple matching entries in `electric-layout-rules' are executed in
order of appearance. When inserting a newline in the 'after-stay
rule, ensure electric-pair-open-newline-between-pairs is nil.
Arguably the logic behind electric-pair-open-newline-between-pairs
should be moved to electric-layout-mode, but the current rule-matching
engine doesn't allow for it. The current solution seems to be good
enough for the situations reported in bug#33794.
* lisp/electric.el (electric-layout-rules): Adjust docstring.
(electric-layout-post-self-insert-function): Loop through rules. Bind
electric-pair-open-newline-between-pairs to nil when handling
after-stay.
Philipp Stephani [Mon, 17 Dec 2018 20:47:46 +0000 (21:47 +0100)]
Add file name handler support for 'make-process' (Bug#28691)
* src/process.c (Fmake_process): Add new keyword argument
':file-handler'.
(syms_of_process) <make-process, :file-handler>: Define new symbols.
* lisp/files.el (file-name-non-special): Add support for
'make-process'.
* test/src/process-tests.el (make-process/file-handler/found)
(make-process/file-handler/not-found)
(make-process/file-handler/disable): New unit tests.
(process-tests--file-handler): New helper function.
* test/lisp/files-tests.el
(files-tests-file-name-non-special-make-process): New unit test.
* doc/lispref/files.texi (Magic File Names): Document that
'make-process' can invoke file name handlers.
* doc/lispref/processes.texi (Asynchronous Processes): Document
':file-handlers' argument to 'make-process'.
* etc/NEWS (Lisp Changes in Emacs 27.1): Mention new
:file-handler argument for 'make-process'.
Macroexpand before evaluating in eval-expression (bis)
* lisp/simple.el (eval-expression): Macroexpand before
evaluating. This repeats the fix made for Bug#20730 in
another branch of the code a few lines down.
Michael Albinus [Fri, 21 Dec 2018 12:53:20 +0000 (13:53 +0100)]
Fix Bug#33735
* lisp/net/tramp-sh.el (tramp-sh-gio-monitor-process-filter):
Delete empty lines. (Bug#33735)
(tramp-sh-inotifywait-process-filter): Use `string-match'.
(tramp-set-remote-path): Check for PIPE_BUF but PATH_MAX.
* test/lisp/filenotify-tests.el: Handle $REMOTE_FILE_NOTIFY_LIBRARY.
(file-notify--deftest-remote): Suppress vc-handled-backends.
(file-notify-test08-backup): Do not expect failure on hydra.
(Bug#33735).
Glenn Morris [Thu, 20 Dec 2018 15:51:17 +0000 (07:51 -0800)]
; Merge from origin/emacs-26
The following commits were skipped:
fda43a7 Skip a vc-bzr test if run as root 4370b8e ; Copyright year fix No need to merge to master 305aca0 Backport: Handle unread-command-events consistently (bug#23980)
Stefan Monnier [Thu, 20 Dec 2018 13:40:43 +0000 (08:40 -0500)]
* lisp/emacs-lisp/map.el: Add support for plists
(map--plist-p, map--plist-delete): New functions.
(map-elt, map-delete, map-length, map-into, map-put!, map-insert)
(map-apply, map-do): Handle the plist case.
* test/lisp/emacs-lisp/map-tests.el (with-maps-do): Add sample plist.
(test-map-put!): The behavior of map-put! is not the same for plists as
for alists.
Alan Mackenzie [Thu, 20 Dec 2018 12:21:16 +0000 (12:21 +0000)]
Check result from c-backward-token-2 to avoid infinite loop
This fixes bug #33784.
* lisp/progmodes/cc-fonts.el (c-get-fontification-context): While moving back
over enclosing parentheses, check that c-backward-token-2 actually moves.
Michael Albinus [Thu, 20 Dec 2018 10:07:15 +0000 (11:07 +0100)]
Fix Bug#33781
* lisp/net/tramp-sh.el (tramp-set-remote-path): Use a temporary
file for setting $PATH, if it exceeds PATH_MAX on the remote system.
(tramp-send-command-and-read): Ignore errors if NOERROR. (Bug#33781)
* test/lisp/net/tramp-tests.el (tramp-test34-remote-path): New test.
Glenn Morris [Wed, 19 Dec 2018 23:37:04 +0000 (18:37 -0500)]
Skip rather than fail in flymake tests
* test/lisp/progmodes/flymake-tests.el
(flymake-tests--wait-for-backends): Skip rather than fail if
backends are slow to respond. This always seems like a temporary
system glitch and not something that should be reported as a failure.
Eric Abrahamsen [Wed, 19 Dec 2018 22:11:14 +0000 (14:11 -0800)]
Make gnus-newsgroup-name local var available to gnus-summary-mode
* lisp/gnus/gnus-sum.el (gnus-summary-mode-group): New defvar for
passing gnus-newsgroup-name (a local variable) into the body of
gnus-summary-mode, so that it's available to gnus-summary-mode-hook.
Paul Eggert [Wed, 19 Dec 2018 21:42:21 +0000 (13:42 -0800)]
cl-make-random-state was not copying its arg
Problem reported by Xu Chunyang (Bug#33731).
* lisp/emacs-lisp/cl-extra.el (cl-make-random-state):
Use copy-sequence, not copy-tree, so that the record is copied.
* test/lisp/emacs-lisp/cl-extra-tests.el:
(cl-extra-test-cl-make-random-state): New test.
Paul Eggert [Wed, 19 Dec 2018 20:57:25 +0000 (12:57 -0800)]
Minor fixes/simplifications to time functions
* doc/lispintro/emacs-lisp-intro.texi (Files List): Simplify.
* doc/lispref/os.texi (Time of Day): Mention format-time-string
as an alternative to current-time-string.
* lisp/arc-mode.el (archive-unixdate, archive-unixtime):
Port better to future versions of Emacs where (COUNT . HZ)
will take precedence to (HI . LO).
* lisp/arc-mode.el (archive-unixtime):
* lisp/calendar/todo-mode.el (todo-insert-item--basic)
(todo-item-done, todo-read-time):
Prefer format-time-string to substringing current-time-string.
* lisp/calc/calc-forms.el (calc-time, calcFunc-now):
Prefer decode-time to parsing the output of current-time-string.
* lisp/emacs-lisp/cl-extra.el (cl--random-time):
Prefer encode-time to hashing the output of current-time-string.
* lisp/gnus/gnus-score.el (gnus-score-headers)
(gnus-score-adaptive):
Avoid stringifying and then reparsing timestamp.
* src/timefns.c (Fencode_time): Omit redundant assignment.
Glenn Morris [Wed, 19 Dec 2018 20:06:18 +0000 (15:06 -0500)]
Skip a vc-bzr test if run as root
* test/lisp/vc/vc-bzr-tests.el (vc-bzr-test-faulty-bzr-autoloads):
Skip this test when run as root. This works around a race
condition in root-specific code in vc-mode-line when deleting a file.
; Do not merge to master, instead fix vc-mode-line there.
* src/keyboard.c (read_char): Events put into `unread-command-events'
with the form (t . EVENT) should always have the t stripped when read
out.
* test/src/keyboard-tests.el: New tests for `unread-command-events'.
Glenn Morris [Wed, 19 Dec 2018 02:18:31 +0000 (21:18 -0500)]
Restrict downcasing in elisp xref tests (bug#25534)
The tests happen to not fail at the moment because find-library-name
now has an extra feature, find-library--from-load-history, which
happens to do a case-insensitive regexp match; but still it seems
better not to rely on this.
* test/lisp/progmodes/elisp-mode-tests.el (xref--case-insensitive):
New variable.
(xref-elisp-test-run, emacs-test-dir): Only downcase if the
filesystem seems to be case-insensitive.
Juri Linkov [Tue, 18 Dec 2018 23:10:09 +0000 (01:10 +0200)]
More font-lock improvements for diff-mode
* lisp/vc/diff-mode.el (diff-font-lock-keywords): Use diff-header face
for git index lines (like already used for bzr index lines).
Use diff-file-header face for binary file headers.
(diff-find-source-location): Use expand-file-name for vc-find-revision.
(diff--font-lock-prettify): Use diff-indicator-* faces for
left-fringe indicators.
(diff-syntax-fontify-props): Optimize to not use text-property-not-all
for font-lock-ensure.
* lisp/replace.el (occur-engine-line): Simplify to use font-lock-ensure
without text-property-not-all.
Juri Linkov [Tue, 18 Dec 2018 22:55:15 +0000 (00:55 +0200)]
Fontify one-line diffs without the final newline (bug#33567)
* lisp/vc/diff-mode.el (diff-hunk-text, diff-syntax-fontify-hunk):
Skip lines beginning with backslash like "\ No newline at end of file".
(diff-syntax-fontify-hunk): Use string-trim-right.
For one-line diffs use 1 explicitly in the list of line numbers.
Rob Browning [Tue, 18 Dec 2018 20:34:01 +0000 (15:34 -0500)]
Avoid test failures if directory name looks like a regexp
Taken from <https://sources.debian.org/patches/emacs/1:26.1+1-1>
* test/lisp/ibuffer-tests.el (ibuffer-filter-inclusion-3):
* test/lisp/net/tramp-tests.el (tramp-test42-remote-load-path):
Regexp-quote file names to avoid failures with directory names
of the form "build/emacs-i87jK3/emacs-26.1+1/...".
Paul Eggert [Tue, 18 Dec 2018 20:21:27 +0000 (12:21 -0800)]
Support (ash INTEGER BIGNUM)
* src/data.c (emacs_mpz_mul_2exp): 2nd arg is now a
nonnegative EMACS_INT not mp_bitcnt_t, to simplify checking.
(Fash): Support COUNT values that are bignums or that
exceed mp_bitcnt_t range.
* test/src/data-tests.el (data-tests-ash-lsh): Test this.
Nicolas Petton [Tue, 18 Dec 2018 08:42:50 +0000 (09:42 +0100)]
New convenience functions in seq.el
Functions to access the first or all but the first elements of
sequences have been repeatedly asked for (the last occurrence being
https://github.com/NicolasPetton/seq.el/issues/9).
* lisp/emacs-lisp/seq.el (seq-first, seq-rest): New functions.
* test/lisp/emacs-lisp/seq-tests.el (test-seq-first, test-seq-rest):
New tests for seq-first and seq-rest.
Stefan Monnier [Mon, 17 Dec 2018 19:51:01 +0000 (14:51 -0500)]
* lisp/emacs-lisp/map.el: Avoid special casing lists.
(map-not-inplace, map-inplace): New errors.
(map-insert): New generic function.
(map-put!): Signal map-not-inplace rather than a generic 'error'.
(map-elt): Use map-not-inplace and map-insert to avoid hardcoding
a special case for lists.
* test/lisp/emacs-lisp/map-tests.el (test-map-put!): Rename from
test-map-put. Also test the errors signaled.
Paul Eggert [Mon, 17 Dec 2018 18:19:23 +0000 (10:19 -0800)]
Some more flatten-tree aliases
* lisp/allout.el (allout-flatten):
* lisp/progmodes/hideif.el (hif-flatten):
Now an obsolete alias for flatten-tree. All callers changed.
* lisp/org/org-protocol.el (org-protocol-flatten):
Make it an alias for flatten-tree if available.
* lisp/progmodes/js.el (js--flatten-list):
Remove alias. We shouldn’t need obsolete function aliases for
private functions.
Paul Eggert [Mon, 17 Dec 2018 17:31:08 +0000 (09:31 -0800)]
Assume ‘emacs’ is defined in Emacs-only code
* src/charset.c, src/coding.c, src/coding.h, src/gmalloc.c:
* src/ralloc.c, src/regex-emacs.c: Simplify slightly by
assuming that ‘emacs’ is defined. These modules have long
been specific to Emacs, and are not used elsewhere.
Paul Eggert [Mon, 17 Dec 2018 16:11:55 +0000 (08:11 -0800)]
More porting to GCC 8 of --enable-gcc-warnings
Backport from master.
I ran into this when building Emacs 26 with GCC 8 on Fedora 29 x86.
* lwlib/lwlib-Xaw.h (xaw_update_one_value, xaw_popup_menu):
* lwlib/lwlib-Xlw.h (xlw_update_one_value, xlw_pop_instance):
* lwlib/lwlib.h (lw_allow_resizing, lw_set_main_areas) [!USE_MOTIF]:
No longer const.
* src/emacs-module.c: Ignore -Wcast-function-type.
João Távora [Sun, 16 Dec 2018 16:55:09 +0000 (16:55 +0000)]
Jsonrpc error handlers can now safely call jsonrpc-shutdown
Previously, if an error handler called jsonrpc-shutdown, and if that
error handler was being called from the process sentinel,
jsonrpc-shutdown would infloop waiting for jsonrpc-sentinel-done to be
set.
Rename the process property jsonrpc-sentinel-done to
jsonrpc-sentinel-cleanup-started, arrange for it to be set earlier
in the sentinel, and also check for it earlier in jsonrpc-shutdown.
* lisp/jsonrpc.el (Version): Bump to 1.0.7.
(jsonrpc--process-sentinel): Set jsonrpc-sentinel-cleanup-started
a bit earlier than previous jsonrpc-sentinel-done.
(jsonrpc-shutdown): Query jsonrpc-sentinel-cleanup-started
Paul Eggert [Sun, 16 Dec 2018 16:11:35 +0000 (08:11 -0800)]
Update from Gnulib
This incorporates:
2018-12-16 regex: propagate fix for glibc bug 18040
2018-12-16 obstack, libc-config: Support HP-UX cc in C99 mode
2018-12-15 regex: work around a bug in glibc-2.27 and prior
2018-12-13 localtime-buffer: Avoid endless recursion
* build-aux/config.guess, build-aux/config.sub, lib/cdefs.h:
* lib/gettext.h, lib/localtime-buffer.c, lib/regexec.c, m4/regex.m4:
Copy from Gnulib.
Stefan Monnier [Sun, 16 Dec 2018 15:32:53 +0000 (10:32 -0500)]
* lisp/net/tramp: Rework mutual dependencies
Functionally split tramp-loaddefs.el into two parts: one part run
while loading it at the very beginning of loading tramp.el (holding plain
autoloads and declarations), and another run at the very end of loading
tramp.el (holding setup code).
This should reduce infinite-recursion problems while loading.
* lisp/net/tramp.el: Require tramp-loaddefs at beginning.
Run (new and internal) tramp--startup-hook at the end.
Remove all tramp-autoloads. Prefer tramp--with-startup over
eval-after-load.
(tramp--with-startup): New macro.
(tramp-set-syntax): Show the new value rather than the old value in the
error message.
(tramp-syntax-values): Explicitly return 'values'.
(tramp-prefix-format, tramp-prefix-regexp, tramp-method-regexp)
(tramp-postfix-method-format, tramp-postfix-method-regexp)
(tramp-prefix-ipv6-format, tramp-prefix-ipv6-regexp)
(tramp-postfix-ipv6-format, tramp-postfix-ipv6-regexp)
(tramp-postfix-host-format, tramp-postfix-host-regexp)
(tramp-remote-file-name-spec-regexp, tramp-file-name-structure)
(tramp-completion-file-name-regexp): Don't initialize in the declaration,
since it's recomputed at the end anyway.
(tramp-time-dont-know, tramp-time-doesnt-exist): Move before first use.
* lisp/net/tramp-compat.el: Don't require tramp-loaddefs any more.
(tramp-compat-tramp-syntax): Declare tramp-syntax.
* lisp/net/tramp-smb.el:
* lisp/net/tramp-rclone.el:
* lisp/net/tramp-gvfs.el:
* lisp/net/tramp-ftp.el:
* lisp/net/tramp-adb.el: Wrap all autoloaded setup operations within
tramp--with-startup.
* lisp/net/tramp-sh.el: Wrap all autoloaded setup operations within
tramp--with-startup.
(tramp-display-escape-sequence-regexp)
(tramp-device-escape-sequence-regexp): Don't tramp-autoload any more.
(tramp-stat-marker): Use eval-and-compile rather than tramp-autoload.
* lisp/net/tramp-sudoedit.el: Wrap all autoloaded setup operations within
tramp--with-startup.
Require tramp-sh at compile-time to precompute some regexps based on
tramp-stat-marker.
* lisp/net/tramp-sh.el (tramp-sh-handle-set-file-uid-gid): Rename from
`tramp-sh-handle-set-file-uid-gid'. Handle only remote file names.
* lisp/net/tramp-sudoedit.el: New file.
* lisp/net/tramp.el (tramp-file-name-for-operation): Handle also
`tramp-set-file-uid-gid'.
(tramp-set-file-uid-gid): New defun.
(tramp-get-local-uid, tramp-get-local-gid): Cache result.
* test/lisp/net/tramp-tests.el (tramp--test-sudoedit-p): New defun.
(tramp-test20-file-modes, tramp-test22-file-times)
(tramp--test-sudoedit-p): Use it.
Glenn Morris [Sat, 15 Dec 2018 23:19:25 +0000 (15:19 -0800)]
Merge from origin/emacs-26
fa953b5 (origin/emacs-26) Fix an epg test for recent GnuPG versions (... 3015636 Document font structure layout constraints f14d574 Document font structure layout constraints f519aa1 Fix display of line numbers in empty lines beyond EOB 24b5026 Fix redisplay when a window's scroll bar or fringes are changed 5d1b158 Tiny markup fix in Elisp manual
(map-empty-p): Add method for lists which avoids computing their
entire length.
(map-contains-key): Check for alist membership by comparing against
DEFAULT argument returned by alist-get.
(map-put!): Reconcile argument name with that used in docstring.
Eli Zaretskii [Fri, 14 Dec 2018 08:43:37 +0000 (10:43 +0200)]
Fix display of line numbers in empty lines beyond EOB
* src/xdisp.c (maybe_produce_line_number): When the current
line is at EOB, use the 'line-number-current-line' face only
on that single line, but not on the rest of empty lines beyond
EOB. (Bug#33732)
Stephen Leake [Thu, 13 Dec 2018 22:45:05 +0000 (14:45 -0800)]
Get long package description for installed packages from installed files
* doc/lispref/package.texi (Archive Web Server): New; document web
server interface.
* lisp/emacs-lisp/package.el (package--get-description): New; get long
description from installed files.
(describe-package-1): Use it, improve comments. No longer writing
NAME-readme.txt.
* test/lisp/emacs-lisp/package-tests.el:
(package-test-describe-package): There is now a description for an
installed package.
(package-test-describe-installed-multi-file-package): New test.
projects / emacs.git / log