* src/alloc.c (make_clear_bool_vector): It’s now the caller’s
responsibility to make sure the bool vector length is in range.
Add an eassert to double-check this. This lets some locals be
ptrdiff_t not EMACS_INT.
(Fmake_bool_vector, Fbool_vector):
Check that bool vector lengths are in range.
* src/lisp.h (BOOL_VECTOR_LENGTH_MAX): New macro.
(bool_vector_words, bool_vector_bytes): Avoid undefined
behavior if size == EMACS_INT_MAX - (BITS_PER_BITS_WORD - 1).
This is mostly theoretical but it’s easy to do it right.
* src/lread.c (read_bool_vector): Use EMACS_INT, not just ptrdiff_t.
Check that length doesn’t exceed BOOL_VECTOR_LENGTH_MAX.
This fixes an unlikely integer overflow where the calculated size
went negative.
* src/fns.c (list_length, Fsafe_length, Fproper_list_p):
Use ptrdiff_t not intptr_t for accumulator, since result is ptrdiff_t.
This fixes a minor glitch in 2019-01-11T05:35:31!eggert@cs.ucla.edu
where I removed unnecessary overflow checks but forgot to change types.
This change should alter generated code only on oddball platforms where
ptrdiff_t is narrower than intptr_t, e.g., CheriBSD.
Previous commit allows Emacs to dump tree-sitter queries by simply
dumping the query string and language symbol, and left the query object
and cursor object as NULL. This commit makes sure Emacs doesn't crash
when loading the dumped query, by make sure Emacs can handle the case
where the cursor is NULL.
* src/treesit.c (make_treesit_query): Initialize query with null cursor.
(treesit_ensure_query_cursor): New function.
(treesit_initialize_query): Ensure cursor is non-null.
* src/treesit.h (Lisp_TS_Query): Update documentation.
* src/comp.c (declare_imported_func, emit_simple_limple_call)
(declare_lex_function, compile_function):
* src/emacs-module.c (funcall_module):
* src/fns.c (Fstring_distance):
* src/font.c (font_sort_entities):
* src/haikumenu.c (digest_menu_items, haiku_menu_show):
* src/pgtkselect.c (Fpgtk_register_dnd_targets):
* src/xfns.c (Fx_begin_drag):
* src/xmenu.c (x_menu_show):
* src/xterm.c (x_dnd_compute_toplevels, handle_one_xevent)
(x_term_init):
Prefer SAFE_NALLOCA to doing size multiplication by hand, to catch
unlikely integer overflows.
* src/comp.c (emit_simple_limple_call): Fix bug where
SAFE_FREE was called too early, leading to unlikely
use of freed storage.
* src/xterm.c (handle_one_xevent): Remove side effects
from SAFE_ALLOCA args, as the args are evaluated twice.
* doc/misc/gnus.texi (Agent Caveats): Change doc due to commit
2020-10-16 "Add a new variable to control Gnus Agent caching"
(41d220dc60, bug#43356). (Bug#72134)
Although loading uninitialized works from memory and then ignoring
the result works fine on conventional architectures, it
technically has undefined behavior in C, so redo bool_vector
allocation so that the code never does that. This can improve
performance when allocating large vectors of nil, since calloc can
clear the memory lazily.
* src/alloc.c (make_clear_bool_vector): New function,
a generalization of make_uninit_bool_vector.
(make_uninit_bool_vector): Use it.
(Fmake_bool_vector): If !INIT, rely on make_clear_bool_vector.
* src/alloc.c (Fbool_vector):
* src/fns.c (Freverse): Don’t access uninitialized bool_vector words.
* lisp/minibuffer.el (completion--hilit-from-re): Avoid signaling
an error if STRING does not match REGEXP. Fix doc string and
indentation. (Bug#72176)
* src/androidgui.h (CONVERT_TO_NATIVE_RECT)
(CONVERT_FROM_EMACS_RECT): Delete redundant macro definitions.
* src/xdisp.c (Fremember_mouse_glyph) [CONVERT_TO_EMACS_RECT]:
Expand CONVERT_TO_EMACS_RECT to convert native rectangles into
a readable format if required. Reported by Stefan Kangas
<stefankangas@gmail.com>.
* lisp/touch-screen.el (touch-screen-current-tool): Expand doc
string.
(touch-screen-handle-point-update): Record extents of glyph
beneath the mouse as computed by `remember_mouse_glyph' if
necessary, and defer generation of mouse-movement events till
the mouse exit it.
* src/xdisp.c (Fremember_mouse_glyph): New function.
(syms_of_xdisp): Define new subr.
* configure.ac (C_SWITCH_MATCHINE): On 32-bit x86 with GCC 4+,
append -mfpmath=sse (if SSE2 is known to work) or -fno-tree-sra
(otherwise) to work around GCC bug 58416.
* etc/NEWS: Mention this.
* INSTALL: Move description of CFLAGS etc. into the section headed
“Here is a complete list of the variables you may want to set”
since they are also variables one might want to set.
Previously, the exit info in Eshell was that of the last command that
finished, rather than the last command in a pipeline.
* lisp/eshell/esh-cmd.el (eshell-exec-lisp)
(eshell-lisp-command): Check whether the command is being piped.
* lisp/eshell/esh-proc.el (eshell-gather-process-output): Record whether
the command is being piped...
(eshell-sentinel): ... and do the right thing with that info.
* test/lisp/eshell/esh-proc-tests.el
(esh-proc-test/sigpipe-exits-process): Check the exit status to ensure
we don't report the first process's SIGPIPE exit.
* lisp/eshell/esh-cmd.el (eshell-last-command-status)
(eshell-last-command-result): Move here from esh-io.el.
(eshell-set-exit-info): New function, extracted from
'eshell-close-handles'.
* lisp/eshell/esh-io.el (eshell-close-handles): Make old calling
convention obsolete. Update callers to use 'eshell-set-exit-info' as
needed.
* lisp/play/gamegrid.el (gamegrid-add-score-insecure): Move point to
the score just added, or end of buffer if the new score did not make
the list. This makes it easier to see where the last game
ranked. (Bug#72185)
* lisp/eshell/esh-proc.el (eshell/kill): Fix handling of commands like
"kill 123". Use REMOTE when signalling PIDs in remote directories.
Signal using process objects when possible. Report errors when failing
to signal.
* test/lisp/eshell/esh-proc-tests.el (esh-proc-test/kill/process-id)
(esh-proc-test/kill/process-object): New tests (bug#72013).
Now, we use the 'eshell-deferrable' wrapper to wrap a form that returns
a process (or list thereof). This improves upon the old method, which
failed to handle 'eshell-replace-command' correctly. In that case,
Eshell would fail to unmark commands as deferrable when necessary
(e.g. for commands in pipelines).
* lisp/eshell/esh-cmd.el (eshell-deferrable-commands): Make into a
defvar.
(eshell-deferrable): New function...
(eshell-structure-basic-command): ... use it.
(eshell-trap-errors): Rename to...
(eshell-do-command): ... this, and use 'eshell-deferrable'. Update
callers.
(eshell--unmark-deferrable): Remove. Update callers.
(eshell-execute-pipeline): Remove 'eshell-process-identity'.
(eshell-process-identity, eshell-named-command*, eshell-lisp-command*):
Make obsolete.
* test/lisp/eshell/esh-cmd-tests.el (eshell-test-replace-command): New
function.
(esh-cmd-test/pipeline/replace-command): New test.
* src/pgtkfns.c (pgtk_get_defaults_value, pgtk_set_defaults_value):
Factor out new function...
(pgtk_check_resource_key_length): ...to here. Avoid unnecessary work by
using strnlen.
* src/insdel.c (del_range_2): Update *_BYTE variables _after_
updating the corresponding character values. This follows what we
do everywhere else, and allows to put a watchpoint on, say, Z_BYTE
to check consistency between the character and byte counts. See
bug#72165 for one situation where it is useful.
