Paul Eggert [Thu, 22 Aug 2019 05:29:35 +0000 (22:29 -0700)]
Remove no-longer-needed workaround for GC bug
* src/keymap.c (describe_vector): Remove old workaround for GC bug.
This workaround, introduced in 1993-02-19T05:43:54Z!rms@gnu.org,
has not been needed for some time. Problem reported by Pip Cet in:
https://lists.gnu.org/r/emacs-devel/2019-08/msg00316.html
Paul Eggert [Thu, 22 Aug 2019 05:19:03 +0000 (22:19 -0700)]
Don’t debug fset by default
This GC bug seems to have been fixed, so the check is no longer
needed in production code. From a suggestion by Pip Cet in:
https://lists.gnu.org/r/emacs-devel/2019-08/msg00316.html
* src/alloc.c (SUSPICIOUS_OBJECT_CHECKING) [!ENABLE_CHECKING]:
Do not define.
(find_suspicious_object_in_range, detect_suspicious_free):
Expand to proper dummy expressions if !SUSPICIOUS_OBJECT_CHECKING.
* src/data.c (Ffset): Convert test to an eassert.
Paul Eggert [Thu, 22 Aug 2019 01:54:08 +0000 (18:54 -0700)]
Fix clrhash bug when hash table needs rehashing
Problem reported by Pip Cet in:
https://lists.gnu.org/r/emacs-devel/2019-08/msg00316.html
* src/fns.c (maybe_resize_hash_table): Prefer ASET to gc_aset
where either will do. Simplify appending of Qunbound values.
Put index_size calculation closer to where it’s needed.
(hash_clear): If hash_rehash_needed_p (h), don’t clear the
nonexistent hash vector. Use memclear to speed up clearing.
* src/lisp.h (HASH_TABLE_SIZE): Check that the size is positive,
and tell that to the compiler.
Paul Eggert [Thu, 22 Aug 2019 00:18:33 +0000 (17:18 -0700)]
Don’t hard-loop on cycles in ‘read’ etc.
Problem for ‘read’ reported by Pip Cet in:
https://lists.gnu.org/r/emacs-devel/2019-08/msg00316.html
* src/fns.c (Frequire): Protect against circular current-load-list.
* src/lread.c (Fget_load_suffixes):
Protect against circular load-suffixes or load-file-rep-suffixes.
(Fload): Protect against circular loads-in-progress.
(openp): Protect against circular PATH and SUFFIXES.
(build_load_history): Protect against circular load-history or
current-load-list.
(readevalloop_eager_expand_eval): Protect against circular SUBFORMS.
(read1): Protect against circular data.
* test/src/lread-tests.el (lread-circular-hash): New test.
* lisp/progmodes/hideif.el (hide-ifdef-mode-prefix-key): Make into
a defcustom since it seems like this is something that should be
user-customisable (bug#8922).
Use `quit-window-hook' in Info instead of having its own command
* doc/misc/info.texi (Help-Q): Info now uses `quit-window'.
* lisp/info.el (info-standalone): Adjust doc string.
(Info-exit): Made into obsolete alias.
(Info-mode-map): Bind "q" to `quit-window'.
(Info-mode-menu): Adjust.
(info-tool-bar-map): Ditto.
(Info-mode): Adjust doc string.
(Info-mode): If Info is standalone, kill Emacs on "q".
Call `quit-window-hook' in the `quit-window' command only
* lisp/window.el (quit-restore-window): Don't run quit-window-hook
here...
(quit-window): ... but here instead. Callers that call the former
programmatically can decide themselves whether to call the hook.
(quit-window-hook): Fix doc string.
Paul Eggert [Wed, 21 Aug 2019 07:06:00 +0000 (00:06 -0700)]
Be more careful about pointers to bignum vals
This uses ‘const’ to be better at catching bugs that
mistakenly attempt to modify a bignum value.
Lisp bignums are supposed to be immutable.
* src/alloc.c (make_pure_bignum):
* src/fns.c (sxhash_bignum):
Accept Lisp_Object instead of struct Lisp_Bignum *, as that’s
simpler now. Caller changed.
* src/bignum.h (bignum_val, xbignum_val): New inline functions.
Prefer them to &i->value and XBIGNUM (i)->value, since they
apply ‘const’ to the result.
* src/timefns.c (lisp_to_timespec): Use mpz_t const *
to point to a bignum value.
Paul Eggert [Wed, 21 Aug 2019 00:34:03 +0000 (17:34 -0700)]
Avoid some excess precision in time arithmetic
* doc/misc/emacs-mime.texi (time-date):
Adjust example to match new behavior.
* etc/NEWS: Mention this.
* lisp/calendar/time-date.el (decoded-time-add)
(decoded-time--alter-second):
Don’t lose underestimate precision of seconds component.
* src/bignum.c (mpz): Grow by 1.
* src/timefns.c (trillion_factor): New function.
(timeform_sub_ps_p): Remove.
(time_arith): Avoid unnecessarily-large hz, by reducing the hz
to a value no worse than the worse hz of the two arguments.
The result is always exact unless an error is signaled.
* test/src/timefns-tests.el (timefns-tests--decode-time):
New function.
(format-time-string-with-zone): Test (decode-time LOOK ZONE t)
resolution as well as its numeric value.
Noam Postavsky [Sun, 4 Aug 2019 00:19:31 +0000 (20:19 -0400)]
Respect global-eldoc-mode in minibuffers (Bug#36886)
* lisp/emacs-lisp/easy-mmode.el (define-globalized-minor-mode): Accept
a BODY parameter.
* doc/lispref/modes.texi (Defining Minor Modes): Document new
parameter.
* etc/NEWS: Announce it.
* lisp/simple.el (read--expression): Move eldoc-mode setup to...
* lisp/emacs-lisp/eldoc.el (eldoc--eval-expression-setup): ... here,
new function.
(global-eldoc-mode): Add or remove it to
eval-expression-minibuffer-setup-hook when enabling or disabling
global-eldoc-mode. This enables eldoc in the minibuffer (solving
Bug#27202), only when global-eldoc-mode is enabled.
Paul Eggert [Tue, 20 Aug 2019 21:02:30 +0000 (14:02 -0700)]
Support larger TIMEs in (time-convert TIME t)
Also, improve the doc to match current behavior.
* doc/lispref/os.texi (Time Conversion): Document that
time-convert signals an error for infinite or NaN args,
and that (time-convert TIME t) is exact otherwise.
Mention float-time as an alternative to time-convert.
(Time Calculations): Document that time-add and time-subtract
are exact and do not decrease HZ below the minimum of their args.
* src/timefns.c (decode_float_time): Don’t signal an error for
floating-point arguments whose base-FLT_RADIX exponent is not less
than DBL_MANT_DIG. Instead, convert them to (TICKS . 1) values.
Use two (instead of three) integer exponent comparisons in the
typical case.
* test/src/timefns-tests.el (time-arith-tests):
Add more floating-point tests, including some tests
that the old code fails.
Michael Albinus [Tue, 20 Aug 2019 19:32:34 +0000 (21:32 +0200)]
Another attempt to fix bug#32645
* test/lisp/autorevert-tests.el ()
(auto-revert-test02-auto-revert-deleted-file):
* test/lisp/filenotify-tests.el (file-notify-test04-autorevert):
Check `file-notify-valid-p', not that the descriptor is nil.
Glenn Morris [Tue, 20 Aug 2019 16:01:55 +0000 (09:01 -0700)]
Merge from origin/emacs-26
0b810eb Fix a typo in char-width-table 3f00db7 Minor update in admin/notes/unicode bcd0115 Fix lisp indent infloop on unfinished strings (Bug#37045) 5f992d1 Improve commentary in composite.el 3a04be2 ; Improve commentary in xdisp.c 15de1d1 Fix markup in dired-x.texi bda7fc7 ; Fix typo in a doc string of speedbar.el 6f57ef9 * src/callproc.c (Fcall_process): Doc fix.
Mauro Aranda [Sat, 10 Aug 2019 13:48:24 +0000 (10:48 -0300)]
Don't display wrong ElDoc information when inside ELisp strings or comments
* lisp/progmodes/elisp-mode.el (elisp--fnsym-in-current-sexp): Since
forward-sexp assumes point is not in a string or comment, avoid
calling it and then checking if point is inside a string, since that
sometimes will fail with awkward results. (Bug#35567)
Output `auto-coding-alist' in `describe-current-coding-system'
* lisp/international/mule-diag.el
(describe-current-coding-system): Also output the contents of
`auto-coding-alist', which take precedence over
`file-coding-system-alist' (bug#9575).
* doc/lispref/windows.texi (Quitting Windows): Mention in.
* lisp/window.el (quit-restore-window): Run the new
`quit-window-hook' before doing anything else (bug#9867).
(quit-window): Note that the hook will be run in the doc string.
* lisp/window.el (quit-window-hook): New variable.
Paul Eggert [Tue, 20 Aug 2019 01:04:56 +0000 (18:04 -0700)]
Fix org-table 65536-second bug
* lisp/org/org-table.el (org-table-message-once-per-second):
Fix bug when clock difference goes past a 65536-second boundary.
Don’t assume particular format for current-time result.
Paul Eggert [Tue, 20 Aug 2019 01:02:59 +0000 (18:02 -0700)]
Fix time-add/time-sub validity checking
* src/timefns.c (time_arith): Check the first arg for
validity even if the second arg is not finite.
* test/src/timefns-tests.el (time-arith-tests): Test this.
* doc/lispref/processes.texi (Asynchronous Processes): Note that input
may read when sending data as well.
(Output from Processes): Note that functions which send data may also
trigger reading from processes.
(Input to Processes, Filter Functions): Note that filter functions may
be called recursively.
Tino Calancha [Mon, 19 Aug 2019 15:32:09 +0000 (17:32 +0200)]
Fix query-replace-regexp undo feature
Ensure that non-regexp strings used with `looking-at' are quoted.
* lisp/replace.el (perform-replace): Quote regexp (Bug#37073).
* test/lisp/replace-tests.el (replace-tests-perform-replace-regexp-flag):
New variable.
(replace-tests-with-undo): Use it.
(query-replace-undo-bug37073): Add tests.
Use lexical-binding.
(erc-add-timestamp): Store the timestamp in a closure placed in
cursor-sensor-functions rather than stashing it in an ad-hoc
`timestamp` property.
(erc-echo-timestamp): Simplify accordingly.
Noam Postavsky [Sun, 18 Aug 2019 22:10:50 +0000 (18:10 -0400)]
Handle more subprocess chunking in M-x man (Bug#36927)
* lisp/man.el (Man-bgproc-filter): Make sure not to chop man sections
by narrowing.
(Man-highlight-references0): Revert previous fix, as it's no longer
needed.
* test/lisp/man-tests.el (man-tests-filter-strings): New function.
(man-bgproc-filter-buttonize-includes): New test.
Paul Eggert [Sun, 18 Aug 2019 19:11:06 +0000 (12:11 -0700)]
Improve bignum_integer static checking
* src/bignum.h (bignum_integer): Now returns pointer-to-const,
to catch trivial mistakes where the caller might try to modify
a Lisp bignum. Lisp bignums are supposed to be immutable.
All callers changed.
Eli Zaretskii [Sun, 18 Aug 2019 16:36:31 +0000 (19:36 +0300)]
Support the new Japanese era name
* admin/unidata/NormalizationTest.txt:
* admin/unidata/UnicodeData.txt: Add U+32FF SQUARE ERA NAME REIWA.
Do not merge to master.
* test/lisp/international/ucs-normalize-tests.el
(ucs-normalize-tests--failing-lines-part1)
(ucs-normalize-tests--failing-lines-part2): Update. Do not
merge to master.
Paul Eggert [Sun, 18 Aug 2019 00:19:13 +0000 (17:19 -0700)]
Fix org-timer-show-remaining-time > 1 hour
* lisp/org/org-timer.el (org-timer-show-remaining-time):
Don’t assume the remaining time is less than one hour.
Simplify. The simplification removes the need for a
decode-time, and fixes a typo I introduced recently.
Make `describe-function' say that disabled functions are disabled
* lisp/help-fns.el (help-fns--disabled): New function (bug#10853).
(help-fns-describe-function-functions): Add it to the list of
function help functions.
Paul Eggert [Sat, 17 Aug 2019 22:39:18 +0000 (15:39 -0700)]
Add FIXMEs for subsecond support
This adds FIXMEs to areas where Lisp code should support
subsecond information in broken-down timestamps.
It also fixes some unnecessary truncation of timestamps, and
ports the code to a hypothetical future Emacs version where
(decode-time) returns subsecond timestamps by default.
* lisp/calc/calc-forms.el (calc-time, math-iso-dt-to-date)
(calcFunc-now):
* lisp/calendar/icalendar.el (icalendar--add-decoded-times):
* lisp/calendar/iso8601.el (iso8601-parse-interval):
Truncate seconds to an integer, and add a FIXME about
subseconds support.
* lisp/calendar/icalendar.el (icalendar--decode-isodatetime)
(icalendar--decode-isoduration):
Add a FIXME about subseconds support.
* lisp/gnus/gnus-delay.el (gnus-delay-article):
Don’t truncate seconds to an integer, as there’s no need
to do that here.
* lisp/gnus/gnus-util.el (gnus-seconds-today)
(gnus-seconds-month, gnus-seconds-year):
* lisp/gnus/message.el (message-make-expires-date):
* lisp/org/org-timer.el (org-timer-show-remaining-time):
* lisp/vc/ediff-mult.el (ediff-format-date):
Truncate seconds to an integer, as that’s what’s wanted here.
* lisp/midnight.el (midnight-next):
Ceiling seconds to an integer, as that’s what wanted here.
Noam Postavsky [Fri, 16 Aug 2019 11:26:40 +0000 (07:26 -0400)]
Fix lisp indent infloop on unfinished strings (Bug#37045)
* lisp/emacs-lisp/lisp-mode.el (lisp-indent-calc-next): Stop trying to
skip over strings if we've hit the end of buffer.
* test/lisp/emacs-lisp/lisp-mode-tests.el
(lisp-indent-unfinished-string): New test.
Paul Eggert [Sat, 17 Aug 2019 10:27:58 +0000 (03:27 -0700)]
Update from Gnulib
This incorporates:
2019-08-17 intprops: port to Oracle Developer Studio 12.6
2019-08-14 intprops: support uchar, ushort _WRAPV dests
* lib/intprops.h: Copy from Gnulib.
Paul Eggert [Sat, 17 Aug 2019 05:09:04 +0000 (22:09 -0700)]
More-compatible subsecond calendrical timestamps
Instead of appending a subseconds member to the result of
‘decode-time’, this keeps the format unchanged unless you give
a new optional argument to ‘decode-time’. Also, the augmented
format now puts the subsecond info in the SECONDS element, so
the total number of elements is unchanged; this is more
compatible with code that expects the traditional 9 elements,
such as ‘(pcase decoded-time (`(,SEC ,MIN ,HOUR ,DAY ,MON
,YEAR ,DOW ,DST ,ZONE) ...) ...)’.
* doc/lispref/os.texi, doc/misc/emacs-mime.texi, etc/NEWS:
* lisp/net/soap-client.el (soap-decode-date-time):
* lisp/simple.el (decoded-time):
Document the new behavior.
* lisp/calendar/icalendar.el (icalendar--decode-isodatetime):
* lisp/calendar/iso8601.el (iso8601-parse)
(iso8601-parse-time, iso8601-parse-duration)
(iso8601--decoded-time):
* lisp/calendar/parse-time.el (parse-time-string):
* lisp/calendar/time-date.el (decoded-time-add)
(decoded-time--alter-second):
* lisp/org/org.el (org-parse-time-string):
* lisp/simple.el (decoded-time):
* src/timefns.c (Fdecode_time, Fencode_time):
* test/lisp/calendar/icalendar-tests.el:
(icalendar--decode-isodatetime):
* test/lisp/calendar/iso8601-tests.el (test-iso8601-date-years)
(test-iso8601-date-dates, test-iso8601-date-obsolete)
(test-iso8601-date-weeks, test-iso8601-date-ordinals)
(test-iso8601-time, test-iso8601-combined)
(test-iso8601-duration, test-iso8601-intervals)
(standard-test-dates, standard-test-time-of-day-fractions)
(standard-test-time-of-day-beginning-of-day)
(standard-test-time-of-day-utc)
(standard-test-time-of-day-zone)
(standard-test-date-and-time-of-day, standard-test-interval):
* test/lisp/calendar/parse-time-tests.el (parse-time-tests):
* test/src/timefns-tests.el (format-time-string-with-zone)
(encode-time-dst-numeric-zone):
Revert recent changes that added a SUBSECS member to
calendrical timestamps, since that component is no longer
present (the info, if any, is now in the SECONDS member).
* lisp/calendar/time-date.el (decoded-time-add)
(decoded-time--alter-second):
Support fractional seconds in the new form. Simplify.
* src/timefns.c (Fdecode_time): Support new arg FORM.
(Fencode_time): Support subsecond resolution.
* test/src/timefns-tests.el (format-time-string-with-zone)
(decode-then-encode-time): Test subsecond calendrical timestamps.
Paul Eggert [Sat, 17 Aug 2019 01:08:23 +0000 (18:08 -0700)]
Subtracting “now” from “now” should yield zero
* src/timefns.c (time_arith): Arrange for (time-subtract nil
nil) to yield 0, to be consistent with (time-equal-p nil nil).
* test/lisp/calendar/time-date-tests.el (test-time-since): New test.
Paul Eggert [Fri, 16 Aug 2019 23:25:02 +0000 (16:25 -0700)]
Fix time-add rounding bug
Without this fix, time arithmetic yielded results that were not
mathematically accurate, even though the exact results were
representable; for example, (time-add 0 1e-13) yielded a timestamp
equal to 0 instead of to 1e-13.
* lisp/timezone.el (timezone-time-from-absolute):
Let time-add do its thing rather than using floating point
internally, which has rounding errors. We now have bignums and so
don’t need floating point to avoid overflow issues.
* src/timefns.c (timeform_sub_ps_p): New function.
(time_arith): If either argument is a float, represent the
result exactly instead of discarding sub-ps info.
* test/lisp/timezone-tests.el (timezone-tests-time-from-absolute):
Don’t assume (HI LO US PS) timestamp format.
* test/src/emacs-module-tests.el (mod-test-add-nanosecond/valid):
Don’t assume that time-add discards sub-ns info.
* test/src/timefns-tests.el (time-rounding-tests):
Add regression test to detect time-add rounding bug.
Mention `next-multiframe-window' when talking about `other-window'
* doc/emacs/windows.texi (Other Window): Mention
the `next-multiframe-window' command here (which is otherwise not
documented in the manual) (bug#12431).
Alex Branham [Fri, 16 Aug 2019 21:00:31 +0000 (14:00 -0700)]
Make checkdoc check cl-lib function docstrings
* lisp/emacs-lisp/checkdoc.el (checkdoc-this-string-valid-engine):
Remove calls to delete-region to avoid deleting final " (bug#26328).
* lisp/emacs-lisp/checkdoc.el (checkdoc--next-docstring)
(checkdoc-defun-info): Include cl-defun, cl-defgeneric,
cl-defmethod.
(checkdoc-this-string-valid-engine): Add cl-lib supported
keywords.
(checkdoc-defun-info): Ensure function parameters are a
"flat" list (bug#37034).
Do not recreate full URL for proxied HTTPS requests (Bug#35969)
* lisp/url/url-http.el (url-http-create-request): Do not recreate
full URL for proxied HTTPS requests.
(url-https-proxy-after-change-function): Do not bind
url-http-proxy to nil before calling url-http-create-request.
(Bug#35969)
Reimplement the `fill-flowed' function to respect space stuffing
* lisp/mail/flow-fill.el (fill-flowed): Reimplement the function
to respect space-stuffing (bug#17190).
* test/lisp/mail/flow-fill-tests.el
(fill-flow-tests-fill-flowed-stuffed): New test.
(fill-flow-tests-fill-flowed-decode): Rename the test so that it
actually runs.
Alex Branham [Wed, 26 Jun 2019 18:59:06 +0000 (13:59 -0500)]
Fix eshell-mode-map initialization
* lisp/eshell/esh-mode.el (eshell-mode-map, eshell-command-map): Set
up normal keymaps and prefix commands rather than re-initializing them
in each eshell buffer
* lisp/eshell/em-cmpl.el (eshell-cmpl-mode-map, eshell-cmpl-mode)
(eshell-cmpl-initialize):
* lisp/eshell/em-hist.el (eshell-hist-mode-map, eshell-hist-mode)
(eshell-hist-initialize):
* lisp/eshell/em-pred.el (eshell-pred-mode-map, eshell-pred-mode)
(eshell-pred-initialize):
* lisp/eshell/em-prompt.el (eshell-prompt-mode-map, eshell-prompt-mode)
(eshell-prompt-initialize):
* lisp/eshell/em-rebind.el (eshell-rebind-mode-map, eshell-rebind-mode)
(eshell-rebind-initialize):
* lisp/eshell/esh-arg.el (eshell-arg-mode-map, eshell-arg-mode)
(eshell-arg-initialize):
* lisp/eshell/esh-proc.el (eshell-proc-mode-map, eshell-proc-mode)
(eshell-proc-initialize):
* lisp/eshell/esh-var.el (eshell-var-mode-map, eshell-var-mode)
(eshell-var-initialize): Create a new minor mode with a keymap and
call it in the module initialization function.
Paul Eggert [Thu, 15 Aug 2019 17:51:03 +0000 (10:51 -0700)]
Fix typeof portability issue with bitfields
Problem reported by Glenn Morris in:
https://lists.gnu.org/r/emacs-devel/2019-08/msg00300.html
* src/lisp.h (lisp_h_make_fixnum): Use typeof (+(n)) instead
of typeof (n), so that it works with compilers that do
not allow typeof to be applied to a bitfield.
Paul Eggert [Thu, 15 Aug 2019 17:40:11 +0000 (10:40 -0700)]
Fix rounding errors with float timestamps
When converting from float to (TICKS . HZ) form, do the
conversion exactly. When converting from (TICKS . HZ) form to
float, round to even precisely. This way, successfully
converting a float to (TICKS . HZ) and back yields a value
numerically equal to the original.
* src/timefns.c (flt_radix_power_size): New constant.
(flt_radix_power): New static var.
(decode_float_time): Convert the exact numeric value rather
than guessing TIMESPEC_HZ resolution.
(s_ns_to_double): Remove; no longer needed.
(frac_to_double): New function.
(decode_ticks_hz): It is now the caller’s responsibility to
pass a valid TICKS and HZ. All callers changed.
Use frac_to_double to round (TICKS . HZ) precisely.
(decode_time_components): When decoding nil, use
decode_ticks_hz since it rounds precisely.
(syms_of_timefns): Initialize flt_radix_power.
* test/src/timefns-tests.el (float-time-precision): New test.
Paul Eggert [Thu, 15 Aug 2019 09:18:06 +0000 (02:18 -0700)]
Debug out-of-range make_fixnum args
With --enable-checking, make_fixnum (N) now checks that N is
in fixnum range. Suggested by Pip Cet in:
https://lists.gnu.org/r/emacs-devel/2019-07/msg00548.html
A new function make_ufixnum (N) is for the rare cases where N
is intended to be unsigned and is in the range 0..INTMASK.
* configure.ac (AC_C_TYPEOF): Add.
(HAVE_STATEMENT_EXPRESSIONS): Resurrect this macro.
* src/fns.c (Frandom, hashfn_eq, hashfn_equal, hashfn_user_defined):
* src/profiler.c (hashfn_profiler):
Use make_ufixnum rather than make_fixum, since the argument is
an unsigned integer in the range 0..INTMASK rather than a signed
integer in the range MOST_NEGATIVE_FIXNUM..MOST_POSITIVE_FIXNUM.
Typically this is for hashes.
* src/lisp.h (lisp_h_make_fixnum_wrap) [USE_LSB_TAG]:
Rename from lisp_h_make_fixnum.
(lisp_h_make_fixnum): Redefine in terms of lisp_h_make_fixnum_wrap.
Check for fixnum overflow on compilers like GCC that
have statement expressions and typeof.
(FIXNUM_OVERFLOW_P): Move up.
(make_fixnum): Check for fixnum overflow.
(make_ufixnum): New function, which checks that the arg
fits into 0..INTMASK range.