Paul Eggert [Thu, 19 Jul 2018 20:29:28 +0000 (13:29 -0700)]
Improve doc for floating point ‘=’ vs ‘eql’
* doc/lispref/numbers.texi (Float Basics, Comparison of Numbers):
Improve documentation of ‘=’ vs ‘eq’, ‘eql’ and ‘equal’
when NaNs and signed zeros are involved.
Add 'font-lock-maximum-decoration' levels for Python
* etc/NEWS: New entry under Python mode.
* lisp/progmodes/python.el (python-font-lock-keywords-level-1)
(python-font-lock-keywords-level-2)
(python-font-lock-keywords-maximum-decoration): New variables
based off the incumbent 'python-font-lock-keywords'.
(python-font-lock-keywords): Change it to a list of the new
symbols, for use in the 'car' of 'font-lock-defaults'.
(python-mode): Set the 'car' of 'font-lock-defaults' to the
value of 'python-font-lock-keywords', instead of the symbol
'python-font-lock-keywords'.
Stefan Monnier [Wed, 18 Jul 2018 14:23:20 +0000 (10:23 -0400)]
* lisp/comint.el: Clean up namespace
(shell-strip-ctrl-m): Mark as obsolete.
(comint-send-invisible): Rename from `send-invisible`.
(send-invisible): Make it an obsolete alias.
* lisp/net/rlogin.el: Adjust accordingly; Use lexical-binding.
* lisp/shell.el: Adjust accordingly.
Fix imenu--generic-function after fix for Bug#32024.
* lisp/imenu.el (imenu--generic-function): Restore returning of nconc
result from the function. Move filtering out empty menus so it is
done before removing dummy element and splicing main element into
index-alist.
Paul Eggert [Wed, 18 Jul 2018 10:16:54 +0000 (03:16 -0700)]
Fix bug with eql etc. on NaNs
Fix a bug where eql, sxhash-eql, memql, and make-hash-table
were not consistent on NaNs. Likewise for equal,
sxhash-equal, member, and make-hash-table. Some of these
functions ignored NaN significands, whereas others treated
them as significant. It's more logical to treat significands
as significant, and this typically makes eql a bit more
efficient on floats, with just one integer comparison instead
of one to three floating-point comparisons.
* doc/lispref/numbers.texi (Float Basics): Document that
NaNs are never numerically equal, but might be eql.
* src/fns.c (WORDS_PER_DOUBLE): Move to top level of this file.
(union double_and_words): Now named, and at the top level of this file.
(same_float): New function.
(Fmemql, Feql, internal_equal, cmpfn_eql): Use it, so that
the corresponding functions treat NaNs consistently.
(sxhash_float): Simplify based on above-mentioned changes.
* test/src/fns-tests.el (fns-tests-equality-nan): New test.
Paul Eggert [Tue, 17 Jul 2018 16:26:39 +0000 (09:26 -0700)]
Update from Gnulib
This causes config.guess to assume support for shell functions,
a safe assumption nowadays.
* build-aux/config.guess, build-aux/config.sub: Copy from Gnulib.
Michael Albinus [Tue, 17 Jul 2018 10:03:43 +0000 (12:03 +0200)]
Add variable main-thread, fix Bug#32169
* doc/lispref/threads.texi (Basic Thread Functions): Add example,
how to propagate signals to the main thread. Describe variable
`main-thread'. Document optional argument CLEANUP of
`thread-last-error'.
Improve description of window configs in 'register-val-describe'
* etc/NEWS: Describe the change.
* lisp/register.el (register-val-describe)
[(window-configuration-p (car val))]: Include the shown
buffers in the return value to make
'register-read-with-preview' more useful. (Bug#30863)
Reject gpg 2.0 for epg configs by default (Bug#23561)
Previously, gpg2 2.0 would be rejected, but the same version installed
as "gpg" would be accepted.
* lisp/epg-config.el (epg-gpg2-minimum-version): New constant.
(epg-config--program-alist) <OpenPGP>: Require a version in 1.4.3..2.0
or 2.1.6+., not just anything above 1.4.3.
(epg-check-configuration): Accept a list of required version
intervals, in addtion to just a single minimum.
Don't test symmetric operations on gpg 2.0 (Bug#23561)
On the Hydra test machines, which have gpg 2.0, the symmetric tests
fail.
* test/lisp/epg-tests.el (with-epg-tests): Pass REQUIRE-PUBLIC-KEY to
`epg-tests-find-usable-gpg-configuration' and call it before
`epg-make-context' so that the latter uses the resulting cached
config.
(epg-tests-find-usable-gpg-configuration): Only allow gpg 2.0 for
symmetric operations. Clear `epg--configurations' and don't pass
NO-CACHE to `epg-find-configuration'.
(epg-tests--config-program-alist): Use copy-tree to avoid modifying
the epg-config--program-alist cons values.
* test/lisp/emacs-lisp/package-tests.el (package-test-signed): Allow
running with gpg 2.0.
* test/lisp/emacs-lisp/package-tests.el (package-test-signed): Update
the expected message. The message was changed in 2018-06-25 "Reformat
package.el message strings for future l10n".
(with-package-test, with-fake-help-buffer): Add debug declarations.
* test/lisp/emacs-lisp/package-tests.el (package-test-signed): Stop
using epg-check-configuration and ignore-errors, they're redundant
because epg-find-configuration already does all that.
* test/lisp/epg-tests.el (epg-tests-find-usable-gpg-configuration):
Remove tracing.
(with-epg-tests): Skip test if no gpg config is found.
(epg-decrypt-1 epg-roundtrip-1, epg-sign-verify-1, epg-sign-verify-2)
(epg-import-1): Don't check gpg configuration at top-level, rely on
`with-epg-tests' instead. Checking the gpg configuration requires a
valid HOME (or GNUPGHOME), which is provided by `with-epg-tests'.
Avoid infloop in redisplay due to faulty mode-line properties
* xdisp.c (safe_set_text_properties): New function.
(display_mode_element): Call Fset_text_properties through
internal_condition_case_n, using safe_set_text_properties as a
wrapper. (Bug#32038)
17ebb6e (origin/emacs-26) Use consistent function names in thread-tes... 1c86229 Fix format error in Faccept_process_output b38b91a Lessen stack consumption in recursive read1 3eb4603 Match w32 paths in grep sans --null hits (Bug#32051) 5cc7c4b Fix previous make-network-process change d6a1b69 Another documentation improvement in flyspell.el 9b49a8e Improve documentation of Flyspell 3744fda Provide feature 'threads ef9025f Save the server alias on reconnect (Bug#29657) db3874b Refer to "proper lists" instead of "true lists" 35e0305 Avoid turning on the global-minor-mode recursively 51bf4e4 Fix Bug#32085
* lisp/emacs-lisp/subr-x.el (string-join): #'-quote function symbol.
(string-trim-left, string-trim-right):
Make better use of substring for minor speedup.
* test/lisp/emacs-lisp/subr-x-tests.el
(subr-x-test-string-trim-left, subr-x-test-string-trim-right)
(subr-x-test-string-remove-prefix)
(subr-x-test-string-remove-suffix): New tests.
For discussion, see thread starting at
https://lists.gnu.org/archive/html/emacs-devel/2018-05/msg00222.html.
* lisp/custom.el: (custom-available-themes): Use directory-files
instead of performing arbitrary wildcard expansion in file names.
(custom-theme--load-path): Document return value.
* test/lisp/custom-tests.el: New file.
(custom-theme--load-path): New test.
Michael Albinus [Fri, 13 Jul 2018 12:28:12 +0000 (14:28 +0200)]
Use consistent function names in thread-tests.el
* test/src/thread-tests.el (threads-call-error, threads-custom)
(threads-errors, threads-sticky-point, threads-signal-early):
Rename, using naming convention to prefix with "threads-".
Match w32 paths in grep sans --null hits (Bug#32051)
* lisp/progmodes/grep.el (grep-regexp-alist): Add an optional part to
match paths starting with C: (other drive letters).
* test/lisp/progmodes/compile-tests.el
(compile-tests--grep-regexp-testcases)
(compile-tests--grep-regexp-tricky-testcases)
(compile-test-grep-regexps): New tests.
(compile--test-error-line): Return `compilation-message'.
* lisp/vc/smerge-mode.el (smerge-refine-regions): Bind
write-region-inhibit-fsync to t. This was reported in
https://github.com/magit/magit/pull/2834 to give a noticable speedup.
Don't skip epg-tests even with gpg 2.0 (Bug#23561)
* test/lisp/epg-tests.el (epg-tests--config-program-alist): New
constant, which allows gpg2 version 2.0+.
(epg-tests-find-usable-gpg-configuration): Pass it to
epg-find-configuration.
* src/process.c (Fmake_network_process): On 2018-07-09 "Explicitly
reject :server and :nowait (Bug#31903)", the sense of the SERVER check
was accidentally reversed so that we ended up looking for the wrong
ADDRESS. Reported by T.V Raman in
<https://lists.gnu.org/archive/html/emacs-devel/2018-07/msg00437.html>.
* lisp/subr.el (zerop): Add commentary explaining why moving the
function's location within the file broke bootstrap in
2018-07-10T23:08:58-07:00!contovob@tcd.ie.
Miciah Masters [Mon, 11 Dec 2017 01:14:09 +0000 (20:14 -0500)]
Save the server alias on reconnect (Bug#29657)
rcirc does not retain the server alias on reconnect. As a result, rcirc
fails to re-use server and channel buffers when an alias is used. Further
problems may ensue when aliases are used to differentiate multiple
connections to the same host, for example when using a single IRC bouncer
or proxy to connect to multiple IRC networks.
Save the server alias when connecting to a server so that reconnect will
retain the alias.
* lisp/net/rcirc.el (rcirc-connect): Include server-alias when setting
rcirc-connection-info.
* doc/lispref/lists.texi (Cons Cells, Building Lists):
* doc/lispref/sequences.texi (Vector Functions): Use the more
popular term "proper", rather than "true", to qualify nil-terminated
lists.
For discussion, see the following emacs-devel subthreads:
https://lists.gnu.org/archive/html/emacs-devel/2018-06/msg00112.html
https://lists.gnu.org/archive/html/emacs-devel/2018-06/msg00138.html
Stefan Monnier [Wed, 11 Jul 2018 02:52:21 +0000 (22:52 -0400)]
* lisp/vc/diff-mode.el: Perform hunk refinement from font-lock
Remove redundant :group arguments.
(diff-font-lock-refine): New var.
(diff--refine-hunk): New function, extracted from diff-refine-hunk.
(diff-refine-hunk): Use it.
(diff--font-lock-refine--refresh): New function.
(diff--font-lock-refined): New function.
(diff-font-lock-keywords): Use it.
Fix infinite recursion in eshell/clear (Bug#31326)
* lisp/eshell/esh-mode.el (eshell/clear): Bind
eshell-input-filter-functions to nil to prevent entries like
eshell-smart-display-setup from causing infinite recursion.
John Shahid [Sat, 23 Jun 2018 15:12:44 +0000 (11:12 -0400)]
Avoid turning on the global-minor-mode recursively
* lisp/emacs-lisp/easy-mmode.el (define-globalized-minor-mode): Clear
the buffer-list inside MODE-enable-in-buffers to avoid enabling the
mode recursively. (Bug#31793)
Noam Postavsky [Sat, 30 Jun 2018 13:14:22 +0000 (09:14 -0400)]
Respect field boundaries in indent-line-to (Bug#32014)
* lisp/indent.el (indent-line-to): Use the back-to-indentation point
as the end-point of whitespace removal, rather than
backward-to-indentation which doesn't respect field boundaries.
* test/lisp/emacs-lisp/lisp-mode-tests.el
(lisp-indent-with-read-only-field): Don't expect to fail.
65889a6d12 Fix bootstrap infloop in GNU/Linux alpha 48efd1c98b Minor fix of a recent documentation change 3302b7cd7f Mention the NSM in the gnutls variable doc strings 40c2ce743b Remove test code from last commit e02d8e29c6 Fix Bug#32084 da5d6dbe39 Fix (length NON-SEQUENCE) documentation
Noam Postavsky [Fri, 29 Jun 2018 23:58:58 +0000 (19:58 -0400)]
Stop using indent-line-to in lisp-indent-line (Bug#32014)
This is partial revert of "Remove ignored argument from
lisp-indent-line", because `indent-line-to' doesn't respect field
boundaries.
* lisp/emacs-lisp/lisp-mode.el (lisp-indent-line): Use delete-region
and indent-to instead of `indent-line-to'.
* test/lisp/emacs-lisp/lisp-mode-tests.el
(lisp-indent-with-read-only-field): Expect to pass.
Don't merge to master, we will fix indent-line-to there instead.
* src/process.c (Fmake_network_process): Explicitly check for and
signal an error when passed both :server and :nowait non-nil. In
Emacs 25, :nowait would be ignored in this case, but as of Emacs 26.1
this gives an error, albeit an unclear one. Also remove obsolete
comment regarding configurations lacking non-blocking mode, the
corresponding code was removed in 2012-11-17 "Assume POSIX 1003.1-1988
or later for fcntl.h."
Michael Albinus [Mon, 9 Jul 2018 14:03:49 +0000 (16:03 +0200)]
Fix Bug#32090
* lisp/files-x.el (connection-local-normalize-criteria): Do not
use PROPERTIES anymore.
(connection-local-get-profiles): Rewrite, in order to accept any
property as optional. (Bug#32090)
(connection-local-set-profiles):
Adapt ´connection-local-normalize-criteria' call.
Paul Eggert [Sun, 8 Jul 2018 17:00:17 +0000 (10:00 -0700)]
Fix etc/HELLO searching in grep.el
* lisp/progmodes/grep.el (grep-compute-defaults):
Search for "^Copyright", not "^English", as the latter is no
longer present in etc/HELLO and the former is more likely to
survive future changes to etc/HELLO (Bug#32093).
Mention the NSM in the gnutls variable doc strings
* gnutls.el (gnutls-algorithm-priority): Mention the Network
Security Manager here since this variable is an obvious place
for people concerned about network security to look.
(gnutls-verify-error): Ditto.
(gnutls-min-prime-bits): Ditto.
* doc/emacs/misc.texi (Network Security): Rearrange the
network-security-protocol-checks documentation and try to explain
more what this all means and what checks are triggered.
Michael Albinus [Sun, 8 Jul 2018 11:02:19 +0000 (13:02 +0200)]
Fix Bug#32084
* test/lisp/net/dbus-tests.el (dbus-test02-register-service-own-bus):
Unset $DISPLAY when calling dbus-launch, in order to avoid
possible X11 authentication errors. (Bug#32084)
Suggested by Eli Zaretskii <eliz@gnu.org> in the following threads:
https://lists.gnu.org/archive/html/emacs-devel/2018-07/msg00171.html
https://lists.gnu.org/archive/html/emacs-devel/2018-07/msg00206.html
* doc/lispref/sequences.texi (Sequence Functions): Mention that
'length' signals a 'wrong-type-argument' also when given a
non-sequencep argument.
a427de9 (origin/emacs-26) Fix bug #11732 3a04e15 Improve documentation of 'emacs-lock-mode' 9d6ca5a * lisp/imenu.el (imenu-generic-expression): Doc fix. (Bug#32... fdd7e7d Improve indexing of 'eval-defun' in ELisp manual 10af989 Fix (length CIRCULAR) documentation 271d1f7 Tramp editorials 4abf94f Clarify and improve doc strings of 'eval-last-sexp' and friends 6cfc7a7 Automate upload of Emacs manuals to gnu.org b73cde5 Fix MH-E mail composition with GNU Mailutils (SF#485) 0dce5e5 Speed up 'replace-buffer-contents' some more 00fdce0 * doc/emacs/docstyle.texi: Avoid messing up the html output.
* lisp/imenu.el (imenu--generic-function): Move point to START
before checking whether the current item is inside a comment
or a string. Remove any empty menus that could have been
added before returning. (Bug#32024)
John Shahid [Sun, 1 Jul 2018 18:48:24 +0000 (14:48 -0400)]
Keep interactive uses of 'recenter' backward compatible (Bug#31325)
* window.c (Frecenter): Change the interactive spec to always pass
a non-nil value to the REDISPLAY argument when called interactively.
* window.el (recenter-top-bottom): Make sure 'recenter's second
argument is non-nil everywhere.
* windows.texi (Textual Scrolling): Update documentation of
'recenter'.
Eli Zaretskii [Sat, 7 Jul 2018 07:45:45 +0000 (10:45 +0300)]
Fix recent change in window.c
* src/window.c (scroll_command): Fix minor inefficiency in
last change: don't call Fset_buffer if the buffer is already set to be
what we want. (Bug#31988)