It was lost when execute-extended-command was reimplemented in Lisp.
* src/keyboard.c (Fset__this_command_keys): New function.
(syms_of_keyboard): Defsubr it.
* lisp/simple.el (execute-extended-command): Set this-command-keys
as novice.el expects. (Bug#25612)
See comments by Stefan Monnier in:
http://lists.gnu.org/archive/html/emacs-devel/2017-02/msg00181.html
and by Eli Zaretskii in:
http://lists.gnu.org/archive/html/emacs-devel/2017-02/msg00207.html
* src/fns.c (internal_equal): Do not bypass check for depth
overflow when tail-recursing via a dotted list tail or an overlay
plist, to avoid a rare infloop.
* src/lisp.h (FOR_EACH_TAIL, FOR_EACH_TAIL_SAFE): Take TAIL as an
arg, and update it at each iteration, rather than have callers
access it.tail. All callers changed.
(FOR_EACH_TAIL): Do not check for dotted lists, as this is now
the caller’s responsibility. All callers changed.
(FOR_EACH_TAIL_CONS): Remove. All callers changed.
(struct for_each_tail_internal.tail): Remove; no longer needed.
(FOR_EACH_TAIL_INTERNAL): Remove dotted arg, and set the tail
arg each time through the loop. All callers changed.
It does not allow a for-loop's control var to be an anonymous struct.
* src/lisp.h (struct for_each_tail_internal): New type.
(FOR_EACH_TAIL_INTERNAL): Use it.
As per Eli Zaretskii (Bug#25606#20). Although these calls to
maybe_quit are unnecessary in practice, Eli was not convinced
that the calls are unnecessary.
* src/lisp.h (FOR_EACH_TAIL, FOR_EACH_TAIL_CONS):
Call maybe_quit every so often.
(FOR_EACH_TAIL_INTERNAL): New arg CHECK_QUIT. All callers changed.
Use macros like FOR_EACH_TAIL instead of maybe_quit to
catch list cycles automatically instead of relying on the
user becoming impatient and typing C-g (Bug#25606).
* src/fns.c (Flength, Fmember, Fmemq, Fmemql, Fassq, Fassoc, Frassq)
(Frassoc, Fdelete, Freverse):
Use FOR_EACH_TAIL instead of maybe_quit.
(Fnreverse): Use simple EQ to check for circular list instead
of rarely_quit, as this suffices in this unusual case.
(Fplist_put, Flax_plist_put, Flax_plist_put):
Use FOR_EACH_TAIL_CONS instead of maybe_quit.
(internal_equal): Use FOR_EACH_TAIL_CONS to check lists, instead
of by-hand tail recursion that did not catch cycles.
* src/fns.c (Fsafe_length, Fplist_get):
* src/xdisp.c (display_mode_element):
Use FOR_EACH_TAIL_SAFE instead of by-hand Floyd’s algorithm.
* src/lisp.h (QUIT_COUNT_HEURISTIC): Remove; no longer needed.
(rarely_quit): Simply count toward USHRT_MAX + 1, since the
fancier versions are no longer needed.
(FOR_EACH_TAIL_CONS, FOR_EACH_TAIL_SAFE)
(FOR_EACH_TAIL_INTERNAL): New macros, the last with definiens
mostly taken from FOR_EACH_TAIL.
(FOR_EACH_TAIL): Rewrite in terms of FOR_EACH_TAIL_INTERNAL.
* src/data.c (circular_list): New function.
* src/lisp.h (FOR_EACH_TAIL): Use Brent’s algorithm and C99 for-loop
decl, to eliminate the need for the args TAIL, TORTOISE and N, and
to speed things up a bit on typical hosts with optimization.
All uses changed (Bug#25605).
alloc.c had bitrotted a bit, and used an undefined symbol
stack_base when Emacs was built on Solaris sparc, leading to
compilation failures. Also, code related to __builtin_unwind_init
was unnecessarily duplicated. Fix the bitrot and remove some
duplication.
* src/alloc.c: Remove uses of GC_SAVE_REGISTERS_ON_STACK, since it
is never defined.
(test_setjmp) [!HAVE___BUILTIN_UNWIND_INIT && GC_SETJMP_WORKS]:
Define a no-op dummy, to simplify use.
(test_setjmp) [!GC_SETJMP_WORKS]: Test setjmp_tested_p here rather
than in the caller, to simplify use.
(stacktop_sentry): New type.
(__builtin_unwind_init) [!HAVE___BUILTIN_UNWIND_INIT]: New macro.
(SET_STACK_TOP_ADDRESS): New macro, containing code that was duplicated.
(flush_stack_call_func, Fgarbage_collect): Use it.
(init_alloc): Omit unnecessary initialization.
After dumping, Emacs need not re-test setjmp.
* src/xdisp.c (partial_line_height): New function.
(try_scrolling):
* src/window.c (window_scroll_pixel_based): Use it for calculating the
pixel scroll margin correctly in a window with partial lines.
* src/xdisp.c (maximum-scroll-margin): New variable.
* lisp/cus-start.el: Make it customizable.
* etc/NEWS: Mention it.
* doc/emacs/display.texi (Auto Scrolling):
* doc/lispref/windows.texi (Textual Scrolling): Document it.
* src/window.c (window_scroll_pixel_based): Use it instead of hardcoding
division by 4 (Bug #5718).
Its effective range needs to be clamped between 0 and (window height /
4), so it's better to have this constraint in a single place.
* src/window.c (window_scroll_margin): New function.
(window_scroll_pixel_based, window_scroll_line_based):
(Frecenter, Fmove_to_window_line):
* src/xdisp.c (try_scrolling, try_cursor_movement):
(redisplay_window, try_window, try_window_id): Use it.
* src/lisp.h (STACK_CONS, AUTO_STRING_WITH_LEN):
Parenthesize compound literals that are function call args.
Although this does not fix any bugs, it is the proper style for
macro parenthesization as it means this code will continue to
work even if make_lisp_ptr is changed to a macro.
Problem noted by Eli Zaretskii in:
http://lists.gnu.org/archive/html/emacs-devel/2017-01/msg00721.html
This patch also fixes some other issues in that report.
* src/lisp.h (incr_rarely_quit): Remove.
All callers changed to use rarely_quit directly.
* src/search.c (freeze_buffer_relocation)
(thaw_buffer_relocation): New functions.
(looking_at_1, fast_looking_at, search_buffer):
Use them to fix bug when quitting when buffers are frozen.
* src/sysdep.c (emacs_intr_read): Rename from emacs_nointr_read.
All uses changed.
This fixes some infinite loops that cannot be quitted out of,
e.g., (defun foo () (nth most-positive-fixnum '#1=(1 . #1#)))
when byte-compiled and when run under X. See:
http://lists.gnu.org/archive/html/emacs-devel/2017-01/msg00577.html
This also attempts to keep the performance improvements I recently
added, as much as possible under the constraint that the infloops
must be caught. In some cases this fixes infloop bugs recently
introduced when I removed immediate_quit.
* src/alloc.c (Fmake_list):
Use rarely_quit, not maybe_quit, for speed in the usual case.
* src/bytecode.c (exec_byte_code):
* src/editfns.c (Fcompare_buffer_substrings):
* src/fns.c (Fnthcdr):
* src/syntax.c (scan_words, skip_chars, skip_syntaxes)
(Fbackward_prefix_chars):
Use rarely_quit so that users can C-g out of long loops.
* src/callproc.c (call_process_cleanup, call_process):
* src/fileio.c (read_non_regular, Finsert_file_contents):
* src/indent.c (compute_motion):
* src/syntax.c (scan_words, Fforward_comment):
Remove now-unnecessary maybe_quit calls.
* src/callproc.c (call_process):
* src/doc.c (get_doc_string, Fsnarf_documentation):
* src/fileio.c (Fcopy_file, read_non_regular, Finsert_file_contents):
* src/lread.c (safe_to_load_version):
* src/sysdep.c (system_process_attributes) [GNU_LINUX]:
Use emacs_read_quit instead of emacs_read in places where
C-g handling is safe.
* src/eval.c (maybe_quit): Move comment here from lisp.h.
* src/fileio.c (Fcopy_file, e_write):
Use emacs_write_quit instead of emacs_write_sig in places where
C-g handling is safe.
* src/filelock.c (create_lock_file): Use emacs_write, not
plain write, as emacs_write no longer has a problem.
(read_lock_data): Use emacs_read, not read, as emacs_read
no longer has a problem.
* src/fns.c (rarely_quit): Move to lisp.h and rename to
incr_rarely_quit. All uses changed..
* src/fns.c (Fmemq, Fmemql, Fassq, Frassq, Fplist_put, Fplist_member):
* src/indent.c (compute_motion):
* src/syntax.c (find_defun_start, back_comment, forw_comment)
(Fforward_comment, scan_lists, scan_sexps_forward):
Use incr_rarely_quit so that users can C-g out of long loops.
* src/fns.c (Fnconc): Move incr_rarely_quit call to within
inner loop, so that it catches C-g there too.
* src/keyboard.c (tty_read_avail_input): Remove commented-out
and now-obsolete code dealing with interrupts.
* src/lisp.h (rarely_quit, incr_rarely_quit): New functions,
the latter moved here from fns.c and renamed from rarely_quit.
(emacs_read_quit, emacs_write_quit): New decls.
* src/search.c (find_newline, search_buffer, find_newline1):
Add maybe_quit to catch C-g.
* src/sysdep.c (get_child_status): Always invoke maybe_quit
if interruptible, so that the caller need not bother.
(emacs_nointr_read, emacs_read_quit, emacs_write_quit):
New functions.
(emacs_read): Rewrite in terms of emacs_nointr_read.
Do not handle C-g or signals; that is now for emacs_read_quit.
(emacs_full_write): Replace PROCESS_SIGNALS two-way arg
with INTERRUPTIBLE three-way arg. All uses changed.
The old code that sets and clears immediate_quit was
ineffective except when Emacs is running in terminal mode, and
has problematic race conditions anyway, so remove it. This
will introduce some hangs when Emacs runs in terminal mode,
and these hangs should be fixed in followup patches.
* src/keyboard.c (immediate_quit): Remove. All uses removed.
* src/alloc.c
(purecopy_hash_table) New function, makes a copy of the given hash
table in pure storage.
Add new struct `pinned_object' and `pinned_objects' linked list for
pinning objects.
(Fpurecopy) Allow purifying hash tables
(purecopy) Pin hash tables that are either weak or not declared with
`:purecopy t`, use purecopy_hash_table otherwise.
(marked_pinned_objects) New function, marks all objects in pinned_objects.
(garbage_collect_1) Use it. Mark all pinned objects before sweeping.
* src/lisp.h Add new field `pure' to struct `Lisp_Hash_Table'.
* src/fns.c: Add `purecopy' parameter to hash tables.
(Fmake_hash_table): Check for a `:purecopy PURECOPY' argument, pass it
to make_hash_table.
(make_hash_table): Add `pure' parameter, set h->pure to it.
(Fclrhash, Fremhash, Fputhash): Enforce that the table is impure with
CHECK_IMPURE.
* src/lread.c: (read1) Parse for `purecopy' parameter while reading
hash tables.
* src/print.c: (print_object) add the `purecopy' parameter while
printing hash tables.
* src/category.c, src/emacs-module.c, src/image.c, src/profiler.c,
src/xterm.c: Use new (make_hash_table).
* lisp/net/eww.el (eww-download): Check accessibility of
eww-download-directory to prevent starting a download that will fail
to write.
* src/fileio.c (Faccess_file): Clarify the use of string argument in
the docstring.
There’s no longer need to have QUIT stand for a slug of C statements.
Use the more-obvious function-call syntax instead.
Also, use true and false when setting immediate_quit.
These changes should not affect the generated machine code.
* src/lisp.h (QUIT): Remove. All uses replaced by maybe_quit.
On some microbenchmarks this lets Emacs run 60% faster on my
platform (AMD Phenom II X4 910e, Fedora 25 x86-64).
* src/atimer.c: Include keyboard.h, for pending_signals.
* src/editfns.c (Fcompare_buffer_substrings):
* src/fns.c (Fnthcdr, Fmemq, Fmemql, Fassq, Frassq, Fplist_put)
(Fnconc, Fplist_member):
Set and clear immediate_quit before and after loop instead of
executing QUIT each time through the loop. This is OK for loops
that affect only locals.
* src/eval.c (process_quit_flag): Now static.
(maybe_quit): New function, containing QUIT’s old body.
* src/fns.c (rarely_quit): New function.
(Fmember, Fassoc, Frassoc, Fdelete, Fnreverse, Freverse)
(Flax_plist_get, Flax_plist_put, internal_equal, Fnconc):
Use it instead of QUIT, for
speed in tight loops that might modify non-locals.
* src/keyboard.h (pending_signals, process_pending_signals):
These belong to keyboard.c, so move them here ...
* src/lisp.h: ... from here.
(QUIT): Redefine in terms of the new maybe_quit function, which
contains this macro’s old definiens. This works well with branch
prediction on processors with return stack buffers, e.g., x86
other than the original Pentium.
* src/gnutls.c (emacs_gnutls_handle_error): Demote the normal
peer-closed-connection "The TLS connection was non-properly
terminated" message to a lower level so that it isn't shown to
the user by default.
Characters having both open paren syntax and comment start syntax were
being detected as open parens even when they should have been part a
comment starter (Bug#24870).
* src/syntax.c (in_2char_comment_start): New function, extracted from
`scan_sexps_forward'.
(scan_sexps_forward): Add check for a 2-char comment starter before the
loop. Inside the loop, do that check after incrementing the 'from'
character index. Move the single char comment syntax cases into the
switch instead of special casing them before.
* test/src/syntax-tests.el (parse-partial-sexp-paren-comments):
(parse-partial-sexp-continue-over-comment-marker): New tests.
* src/eval.c (FletX): Report an error for invalid constructs like
‘(let* (a . 0))’, so that ‘let*’ is more consistent with ‘let’.
(lambda_arity): Use plain CHECK_CONS.
* src/fns.c (CHECK_LIST_END): Move from here to lisp.h.
(Fcopy_alist): Remove unnecessary CHECK_LIST call, since
concat does that for us.
(Fnthcdr, Fmember, Fmemql, Fdelete, Fnreverse):
Use CHECK_LIST_END, not CHECK_LIST_CONS. This hoists a
runtime check out of the loop.
(Fmemq): Simplify and use CHECK_LIST_END instead of CHECK_LIST.
(Fassq, Fassoc, Frassq, Frassoc):
Simplify and use CHECK_LIST_END instead of CAR.
(assq_no_quit, assoc_no_quit): Simplify and assume proper list.
(Fnconc): Use plain CHECK_CONS, and do-while instead of while loop.
* src/fontset.c (Fnew_fontset):
* src/frame.c (Fmodify_frame_parameters):
Use CHECK_LIST_END at end, rather than CHECK_LIST at start, for a
more-complete check.
* src/gfilenotify.c (Fgfile_add_watch):
Omit unnecessary CHECK_LIST, since Fmember does that for us.
* src/lisp.h (lisp_h_CHECK_LIST_CONS, CHECK_LIST_CONS):
Remove; no longer used.
(CHECK_LIST_END): New inline function.
* src/xterm.c (x_make_frame_visible): Remove code that waits for the
frame to become visible. We have to deal with invisible frames anyway,
the loop could sometimes before the frame turned visible, and for some
window managers (e.g., XMonad, i3wm) it caused Emacs to get stuck in a
busy loop (Bug#24091).
'let' and 'let*' document that their first argument has to be a list,
but don't check for that; instead, they allow (and silently ignore)
other types. Introduce an explicit type check.
* src/eval.c (Flet, FletX): Check that the variable list is indeed a
list.
* test/src/eval-tests.el: Add unit tests.
* src/thread.c (lisp_mutex_lock_for_thread): New function,
with all the guts of lisp_mutex_lock.
(lisp_mutex_lock): Call lisp_mutex_lock_for_thread.
(condition_wait_callback): Don't call post_acquire_global_lock
before locking the mutex, as that could cause a signaled thread to
exit prematurely, because the condvar's mutex is recorded to be
not owned by any thread, and with-mutex wants to unlock it as part
of unwinding the stack in response to the signal.
* src/thread.c (last_thread_error): New static variable.
(syms_of_threads): Staticpro it.
(record_thread_error, Fthread_last_error): New functions.
(syms_of_threads): Defsubr Fthread_last_error.
* doc/lispref/threads.texi (Basic Thread Functions): Document
thread-last-error.
* test/src/thread-tests.el (thread-errors, thread-signal-early)
(threads-condvar-wait): Test the values returned by
thread-last-error.
* src/thread.c (lisp_mutex_lock, lisp_mutex_unlock)
(lisp_mutex_unlock_for_wait, condition_wait_callback)
(condition_notify_callback): Improve commentary.
(condition_wait_callback): Call post_acquire_global_lock before
attempting to lock the mutex, to make sure the lock's owner is
recorded correctly.
* test/src/thread-tests.el (threads-condvar-wait): New test.
While the stack is increased in main(), to allow the regex stack
allocation to use alloca we also need to modify regex.c to actually take
advantage of the increased stack, and not limit stack allocations to
SAFE_ALLOCA bytes.
* src/regex.c (MATCH_MAY_ALLOCATE): Remove obsolete comment about
allocations in signal handlers which no longer happens and correct
description about when and why MATCH_MAY_ALLOCATE should be defined.
(emacs_re_safe_alloca): New variable.
(REGEX_USE_SAFE_ALLOCA): Use it as the limit of stack allocation instead
of MAX_ALLOCA.
(emacs_re_max_failures): Rename from `re_max_failures' to avoid
confusion with glibc's `re_max_failures'.
* src/emacs.c (main): Increase the amount of fixed 'extra' bytes we add
to the stack. Instead of changing emacs_re_max_failures based on the
new stack size, just change emacs_re_safe_alloca; emacs_re_max_failures
remains constant regardless, since if we run out stack space SAFE_ALLOCA
will fall back to heap allocation.
Co-authored-by: Eli Zaretskii <eliz@gnu.org>
The regex stack limit was being computed as the number of stack entries,
whereas it was being compared with the current size as measured in
bytes. This could cause indefinite looping when nearing the stack limit
if re_max_failures happened not to be a multiple of sizeof
fail_stack_elt_t (Bug #24751).
* src/regex.c (GROW_FAIL_STACK): Compute both current stack size and
limit as numbers of stack entries.
