Problem reported by Paul Pogonyshev (Bug#33034).
* src/lisp.h (union specbinding): New member unwind.eval_depth.
* src/eval.c (record_unwind_protect, set_unwind_protect): Set it.
(do_one_unbind): Use it.
* src/json.c (struct json_buffer_and_size): New member
inserted_bytes.
(json_insert): Instead of creating a string and inserting it
into the current buffer, copy the unibyte text into the gap.
(struct json_insert_data): New member inserted_bytes.
(json_insert_callback): Update commentary. Pass the
inserted_bytes value to json_insert and on its return copy the
updated value back into DATA.
(Fjson_insert): Decode the unibyte text inserted into the gap.
Call before-change-functions and after-change-functions only
once, before and after processing the insertion of the entire
JSON representation.
* test/src/json-tests.el (json-insert/throw): Adapt to the
modified implementation of json-insert: it no longer calls the
modification hooks once for each inserted chunk of JSON
representation.
* src/bignum.c (make_neg_biguint): New function.
* src/lread.c (read1): Do not mishandle an unquoted symbol
with name equal to something like "1\0x", i.e., a string
of numeric form followed by a NUL byte.
Formerly these symbols were misread as numbers.
(string_to_number): Change last argument from an integer flag
to a pointer to the length. This lets the caller figure out
how much of the prefix was used. All callers changed.
Add a fast path if the integer (sans sign) fits in uintmax_t.
Update comments and simplify now that bignums are present.
* src/print.c (print_object): Fix quoting of symbols that look
like numbers, by relying on string_to_number for the tricky
cases rather than trying to redo its logic, incorrectly. For
example, (read (prin1-to-string '\1e+NaN)) formerly returned
"1e+NaN", which was wrong: a backslash is needed in the output
to prevent it from being read as a NaN. Escape NO_BREAK_SPACE
too, since lread.c treats it like SPACE.
* test/src/print-tests.el (print-read-roundtrip):
Add tests illustrating the abovementioned bugs.
14c032d Avoid assertion violations in nonsensical calls to 'signal'
b99192f * lisp/simple.el (transient-mark-mode): Correct documentation...
7e42294 Update the locale and language database
8c53d9f Fix a typo in a doc string.
79bda3b Make nneething allow CRLF-encoded files (bug#32940)
This bug on 32-bit platforms was caused by the timespec_hz
definition going haywire because the C expression
FIXNUM_OVERFLOW_P (MOST_POSITIVE_FIXNUM) did not work in #if.
Eventually the numeric problem showed up as a malfunctioning
cursor (Bug#32992). Fix the problem with MOST_POSITIVE_FIXNUM.
By the way, make_fixnum should check for integer overflow when
debugging; this would have made it easier to track this bug down.
But one fix at a time.
* src/lisp.h (INTTYPEBITS): Now a macro, so usable in #if.
(MOST_POSITIVE_FIXNUM): Mention it’s used in #if.
Backport from master.
* src/scroll.c (struct matrix_elt): Change unsigned char fields to
int to handle frames with more than 255 lines (Bug#32951).
Copyright-paperwork-exempt: yes
* src/eval.c (Fsignal): If both arguments are nil, replace the
first one with 'error', to avoid assertion violations further
down the line. (Bug#32961)
* src/scroll.c (struct matrix_elt): Change unsigned char fields to
int to handle frames with more than 255 lines (Bug#32951).
Copyright-paperwork-exempt: yes
This patch is in response to Eli's review (Bug#32902#10).
* src/systime.c: Doc strings of affected functions now refer
to format-time-string instead of to Lisp manual, and
format-time-string's doc string covers time values.
* test/src/systime-tests.el (format-time-string-with-zone):
Check decode-time too.
(decode-then-encode-time, time-arith-tests): New tests.
This follows on a suggestion by Stefan Monnier in:
https://lists.gnu.org/r/emacs-devel/2018-08/msg00991.html
(Bug#32902).
* doc/lispref/buffers.texi (Modification Time):
* doc/lispref/os.texi (Processor Run Time, Time Calculations)
* doc/lispref/processes.texi (System Processes):
* doc/lispref/text.texi (Undo):
Let the "Time of Day" section cover timestamp format details.
* doc/lispref/os.texi (Time of Day):
Say that timestamp internal format should not be assumed.
Document new (ticks . hz) format. Omit mention of seconds-to-time
since it is now just an alias for encode-time.
(Time Conversion): Document encode-time extension.
* etc/NEWS: Mention changes.
* lisp/calendar/cal-dst.el (calendar-system-time-basis): Now const.
* lisp/calendar/cal-dst.el (calendar-absolute-from-time)
(calendar-time-from-absolute)
(calendar-next-time-zone-transition):
* lisp/emacs-lisp/timer.el (timer-next-integral-multiple-of-time):
Simplify by using bignums, (TICKS . HZ), and new encode-time.
* lisp/emacs-lisp/timer.el (timer-next-integral-multiple-of-time):
Simplify by using bignums and new encode-time.
* lisp/calendar/parse-time.el (parse-iso8601-time-string):
Handle DST more accurately, by using new encode-time.
* lisp/calendar/time-date.el (seconds-to-time):
* lisp/calendar/timeclock.el (timeclock-seconds-to-time):
Now just an alias for encode-time.
* lisp/calendar/time-date.el (days-to-time):
* lisp/emacs-lisp/timer.el (timer--time-setter):
* lisp/net/ntlm.el (ntlm-compute-timestamp):
* lisp/obsolete/vc-arch.el (vc-arch-add-tagline):
* lisp/org/org-id.el (org-id-uuid, org-id-time-to-b36):
* lisp/tar-mode (tar-octal-time):
Don't assume timestamps default to list form.
* lisp/tar-mode.el (tar-parse-octal-long-integer):
Now an obsolete alias for tar-parse-octal-integer.
* src/keyboard.c (decode_timer): Adjust to changes to
time decoding functions elsewhere.
* src/timefns.c: Include bignum.h, limits.h.
(FASTER_TIMEFNS): New macro.
(WARN_OBSOLETE_TIMESTAMPS, CURRENT_TIME_LIST)
(timespec_hz, trillion, ztrillion):
New constants.
(make_timeval): Use TIME_T_MAX instead of its definiens.
(check_time_validity, time_add, time_subtract):
Remove. All uses removed.
(disassemble_lisp_time): Remove; old code now folded into
decode_lisp_time. All callers changed.
(invalid_hz, s_ns_to_double, ticks_hz_list4, mpz_set_time)
(timespec_mpz, timespec_ticks, time_hz_ticks)
(lisp_time_hz_ticks, lisp_time_seconds)
(time_form_stamp, lisp_time_form_stamp, decode_ticks_hz)
(decode_lisp_time, mpz_time, list4_to_timespec):
New functions.
(decode_float_time, decode_time_components, lisp_to_timespec):
Adjust to new struct lisp_time, which does not lose
information like the old one did.
(enum timeform): New enum.
(decode_time_components): New arg FORM. All callers changed.
RESULT and DRESULT are now mutually exclusive; no callers need
to change because of this.
(decode_time_components, lisp_time_struct)
(lisp_seconds_argument, time_arith, make_lisp_time, Ffloat_time)
(Fencode_time):
Add support for (TICKS . HZ) form.
(DECODE_SECS_ONLY): New constant.
(lisp_time_struct): 2nd arg is now enum timeform, not int.
All callers changed.
(check_tm_member): Support bignums.m
(Fencode_time): Add new two-arg functionality.
* src/systime.h (struct lisp_time): Now ticks+hz rather than
hi+lo+us+ps, since ticks+hz does not lose info.
* test/src/systime-tests.el (time-equal-p-nil-nil):
New test.
This refactoring will help improve timestamp handling later
(Bug#32902).
* src/bignum.c (mpz_set_uintmax): Move to bignum.h,
and make inline.
(mpz_set_uintmax_slow): Now extern.
(mpz_to_intmax, mpz_to_uintmax): New functions, with
implementation taken from the old bignum_to_intmax
and bignum_to_uintmax.
(bignum_to_intmax, bignum_to_uintmax): Use them.
This does not change behavior; it’s just long-overdue
refactoring (Bug#32902).
* src/emacs.c (main): Call init_timefns, syms_of_timefns.
* src/timefns.c: New file, containing timestamp-related stuff
from editfns.c and sysdep.c.
* src/Makefile.in (base_obj): Add timefns.o.
* src/editfns.c: Simplify by moving a big chunk to timefns.c.
Do not include systime.h, sys/resource.h, sys/param.h,
strftime.h, coding.h.
(HAVE_TZALLOC_BUG, TM_YEAR_BASE, HAVE_TM_GMTOFF, tzeqlen)
(local_tz, utc_tz, emacs_localtime_rz, emacs_mktime_z)
(invalid_time_zone_specification, xtzfree, tzlookup)
(TIME_T_MIN, TIME_T_MAX, time_overflow, invalid_time)
(check_time_validity, hi_time, lo_time, Fcurrent_time)
(time_add, time_subtract, time_arith, Ftime_add)
(Ftime_subtract, Ftime_less_p, Fget_internal_run_time)
(make_lisp_time, disassemble_lisp_time, decode_float_time)
(lisp_to_timespec, lisp_time_struct, lisp_time_argument)
(lisp_seconds_argument, Ffloat_time, emacs_nmemftime)
(Fformat_time_string, format_time_string, Fdecode_time)
(check_tm_member, Fencode_time, Fcurrent_time_string)
(tm_gmtoff, Fcurrent_time_zone, Fset_time_zone_rule)
(emacs_getenv_TZ, emacs_setenv_TZ): Move to timefns.c.
* src/emacs.c (main): Adjust to initialization changes.
* src/sysdep.c: Include <sys/resource.h> if it's present.
Regularize includes a bit.
(Fget_internal_run_time): Move here from editfns.c.
(init_timefns, syms_of_timefns): New functions.
* src/w32.h (w32_get_internal_run_time): Move decl here
so that it need not be cloned.
* test/src/editfns-tests.el:
* test/src/editfns-tests.el (format-time-string-with-zone)
(format-time-string-with-outlandish-zone)
(editfns-tests--have-leap-seconds)
(format-time-string-with-bignum-on-32-bit):
Move to ...
* test/src/timefns-tests.el: ... this new file.
* src/minibuf.c (syms_of_minibuf) <Vread_hide_char>: Clarify
documentation and mention where else the variable is used.
* doc/lispref/minibuf.texi (Reading a Password): Add an index
entry for 'read-hide-char'.
7296b6f Improve cl-do, cl-do* docstrings
d416109 Avoid returning early in 'while-no-input' due to subprocesses
e8a4d94 Cleanup when opening a new terminal fails. (Bug#32794)
# Conflicts:
# etc/NEWS
* src/xdisp.c (cursor_row_fully_visible_p): Handle the case of
make-cursor-line-fully-visible being a function. Accept a 3rd
argument; if non-zero, assume the caller already tested the
conditions for the cursor being fully-visible, and don't
recheck them. All callers changed.
(try_cursor_movement, try_window_id): Call
cursor_row_fully_visible_p instead of testing the value of
make-cursor-line-fully-visible directly.
(syms_of_xdisp) <make-cursor-line-fully-visible>: Update the
doc string. Define a symbol Qmake_cursor_line_fully_visible.
(Bug#32848)
* lisp/cus-start.el (standard): Update the Custom form.
* etc/NEWS: Mention the change in possible values of
'make-cursor-line-fully-visible'.
* src/keyboard.c (kbd_buffer_store_buffered_event): Support
also the internal buffer-switch events.
(syms_of_keyboard) <Qbuffer_switch>: New DEFSYM.
* lisp/subr.el (while-no-input-ignore-events): Ignore
'buffer-switch' events. Reported by Michael Heerdegen
<michael_heerdegen@web.de>.
* etc/NEWS: Mention the change in behavior of 'while-no-input'
* src/term.c (init_tty): Call delete_terminal_internal if emacs_open
fail.
* src/terminal.c (delete_terminal): Move some code into
delete_terminal_internal and call it.
(delete_terminal_internal): New function.
* src/termhooks.h: Prototype for delete_terminal_internal.
See bug#31904 and bug#32812.
* src/nsterm.m (ns_update_begin): Don't lock focus, only clip if there
is already a view focused.
(ns_update_end): Don't mess with view focusing any more.
(ns_focus): Only clip drawing if there is already a focused view,
otherwise mark area dirty for later drawing. Renamed ns_clip_to_rect.
All callers changed.
(ns_unfocus): Don't unfocus the view any more. Renamed
ns_reset_clipping. All callers changed.
(ns_clip_to_row): Update to match ns_clip_to_rect.
(ns_clear_frame):
(ns_clear_frame_area):
(ns_draw_fringe_bitmap):
(ns_draw_window_cursor):
(ns_draw_vertical_window_border):
(ns_draw_window_divider):
(ns_dumpglyphs_stretch):
(ns_draw_glyph_string): Only draw if ns_focus or ns_clip_to_row
return YES.
(ns_copy_bits): Remove superfluous calls to ns_(un)focus.
(ns_flush_display): New function.
This is for consistency with time-less-p.
* doc/lispref/os.texi (Time Calculations), etc/NEWS:
* src/editfns.c (Ftime_equal_p, syms_of_editfns):
* test/lisp/emacs-lisp/timer-tests.el (timer-test-multiple-of-time):
Rename.
* doc/lispref/os.texi (Time Calculations):
Document time-equal, and the behavior on NaNs and infinities of
time-less-p, time-add, time-subtract.
* etc/NEWS: Mention the change.
* src/editfns.c (time_arith): Change last arg from function
to bool. All callers changed. Do the right thing with
infinities and NaNs.
(time_cmp): New function, which handlesx infinities and NaNs.
(Ftime_less_p): Use it.
(Ftime_equal): New function.
* test/lisp/emacs-lisp/timer-tests.el (timer-test-multiple-of-time):
Use it.
* src/bignum.c (mpz_bufsize): New function.
(bignum_bufsize): Use it.
(mpz_get_d_rounded): New function.
(bignum_to_double): Use it.
* src/bignum.c (bignum_to_double):
* src/data.c (bignum_arith_driver):
When converting bignums to double, round instead of
truncating, to be consistent with what happens with fixnums.
* test/src/floatfns-tests.el (bignum-to-float): Test rounding.
Formerly nil meant both that DST was not in effect and that
the DST flag was unknown, and different functions interpreted
the flag differently. Now the meaning is consistently nil for
DST not in effect, and -1 for DST flag not known.
* doc/lispref/os.texi (Time Conversion): The DST slot is
now three-valued, not two-.
* doc/misc/emacs-mime.texi (time-date): Adjust to new behavior.
* etc/NEWS: Mention this.
* lisp/calendar/parse-time.el (parse-time-string):
* src/editfns.c (Fdecode_time):
Return -1 for unknown DST flag.
* test/lisp/calendar/parse-time-tests.el (parse-time-tests):
Adjust tests to match new behavior, and add a new
test for nil vs -1.
d28d54c (origin/emacs-26) More accurate docs for 'text-char-description'
b3baf99 Document synchronous behavior of eshell/make (Bug#32513)
98544ea Fix bs-show with wide characters (Bug#17822)
85af51b Improve Custom menu labels for 2 options
72a2a36 Improve wording of last change in dired-x.texi
d4fa83b Fix GnuTLS test suite with GnuTLS versions 3.4.x
b5bee6b Fix build with gnutls versions 3.0 to 3.2 (Bug#32446)
67eb80e ; * etc/enriched.txt (hanging-indents): Remove extra indent.
c71cfb7 Fix the Bubbles game on TTY frames
3bbf21b Add choice to reshow certificate information (Bug#31877)
6f2c471 * src/alloc.c (Fbool_vector, Flist, Fvector): Doc tweak.
39eecb3 * src/alloc.c (vector): Fix grammatical error in doc string: ...
* src/keymap.c (Ftext_char_description):
* doc/lispref/help.texi (Describing Characters): More accurate
description of 'text-char-description'. Remove incorrect
examples from the ELisp manual. (Bug#32743)
* src/gnutls.c (gnutls_cipher_get_tag_size): Make it return
zero only for versions of GnuTLS < 3.2.2, where
gnutls_cipher_get_tag_size was introduced. This fixes the
GnuTLS test suite, which assumes that any cipher whose tag
size is non-zero is AEAD-capable, and doesn't test such ciphers
if AEAD is not available, i.e. for GnuTLS < 3.5.1. (Bug#32446)
We previously used functions available only in 3.2+ for all 3.x
versions.
* src/gnutls.c [GNUTLS_VERSION_NUMBER < 0x030501]: Replace calls to
gnutls_cipher_get_tag_size with 0.
[GNUTLS_VERSION_NUMBER < 0x030200]: Alias gnutls_cipher_get_iv_size
to gnutls_cipher_get_block_size, gnutls_digest_list to
gnutls_mac_list, and gnutls_digest_get_name to gnutls_mac_get_name.
[WINDOWSNT]: Adjust DLL function definitions and declarations
accordingly.
This change was motivated by the desire to remove the weird
dependency of lib-src/profile.o on src/systime.h. profile.c
included systime.h only for current_timespec, and this
inclusion required systime.h to have #ifdef emacs in multiple
places and complicated further changes I have in mind.
The current_timespec decl belongs in timespec.h anyway,
and the main effect of this change is to move it there.
* lib-src/profile.c (INLINE): Remove.
Include timespec.h, not systime.h.
* lib/gettime.c (gettime): Prefer clock_gettime to nanotime,
and don’t worry about it failing on a CLOCK_REALTIME arg.
POSIX requires it to succeed and I don’t know of any
counterexamples where the fallbacks would work.
(current_timespec): Move here from src/systime.h.
Nowadays it seems to be better to not have this function
be inline.
* lib/timespec.h: Include arg-nonnull.h.
(current_timespec): New declaration.
(gettime, settime): Declare args to be nonnull.
* lib/gettime.c, lib/timespec.h: Copy from Gnulib.
* src/systime.h: Simplify by assuming ‘emacs’ is defined,
which it always is now.
(current_timespec): Move to lib/timespec.h.
* src/atimer.c (start_atimer, debug_timer_callback):
Don’t assume support for setitimer merely because struct
itimerspec works. POSIX no longer requires support for the
obsolescent setitimer function.
