* src/sqlite.c (sqlite3_errstr) [WINDOWSNT]: Define and load from
the DLL only if the SQLite3 version is at least 3.7.15.
(sqlite_prepare_errdata) [SQLITE_VERSION_NUMBER >= 3007015]: Use
the original code if sqlite3_errstr is available.
* src/xterm.c (xi_populate_device_from_info): Use xnmalloc.
Avoid reading classes at all when the XI library only supports
2.0.
(xi_handle_new_classes): New function.
(xi_handle_device_changed): Move class parsing logic there to
avoid duplicating code.
This is for gcc 12.2.1 20220819 (Red Hat 12.2.1-2) x86-64
when Emacs is configured with --enable-gcc-warnings.
* src/buffer.c (Fmove_overlay): Prefer BASE_EQ to EQ in a place
where they’re equivalent because the only symbol allowed here is nil.
* src/xterm.c (xi_has_scroll_valuators): New function.
(xi_handle_device_changed): If the device changed event provides
scroll valuators, then use the values in there. (bug#58980)
The most obvious problem was the lack of support for
`insert-before-markers`, but the behavior was also different in a few
other cases.
* src/itree.h (itree_insert_gap):
* src/itree.c (itree_insert_gap): Add `before_markers` arg.
* src/lisp.h (adjust_overlays_for_insert):
* src/buffer.c (adjust_overlays_for_insert): Add `before_markers` arg.
* src/insdel.c (adjust_markers_for_replace, adjust_markers_for_insert)
(adjust_markers_for_delete): Adjust overlays directly from here.
(insert_1_both, insert_from_string_1, insert_from_gap)
(insert_from_buffer_1, adjust_after_replace, replace_range)
(replace_range_2, del_range_2): Don't adjust overlays explicitly here
any more.
* test/src/buffer-tests.el (test-overlay-insert-before-markers-empty)
(test-overlay-insert-before-markers-non-empty): New tests.
Some names under the interval_* namespace were renamed under the
itree_* namespace in commits:
0. f421b58db5 of 2022-10-19
"Prefix all itree.h type names with itree_".
1. 37a1145410 of 2022-10-19
"Rename all exported itree.h functions with the itree_ prefix"
Further, some values still referenced in commentary were removed in
commits:
2. 258e618364 of 2022-10-17
"Delete the itree_null sentinel node, use NULL everywhere."
3. 2c4a3910b3 of 2022-10-02
"itree: Use a single iterator object"
* src/emacs.c (main): Allocate global itree iterator once and for
all.
* src/alloc.c (mark_overlay):
* src/buffer.c (set_overlays_multibyte):
* src/itree.c (itree_destroy): Update commentary.
(interval_stack_ensure_space, itree_insert_gap): Prefer
unsigned-to-unsigned comparisons over signed-to-unsigned.
(interval_stack_push_flagged, interval_tree_insert)
(interval_tree_contains, itree_iterator_start)
(itree_iterator_finish, itree_iterator_next, itree_iterator_narrow):
Improve assertions.
(itree_init): Rename...
(init_itree): ...to this, for consistency with other global init
functions.
(itree_create): Stop leaking a global iterator allocation on each
call.
(interval_tree_init): Complete renames of
interval_tree -> itree_tree and interval_tree_clear -> itree_clear.
(interval_tree_remove_fix): Fix indentation.
* src/itree.h: Declare init_itree.
(ITREE_FOREACH): Fix typo in commentary.
* src/pdumper.c [CHECK_STRUCTS]
(dump_interval_node): Use the correct name in the HASH condition
and #error message.
(dump_overlay, dump_buffer): Update HASH (bug#58975).
* src/xterm.c (xi_reset_scroll_valuators_for_device_id): Minor
style adjustments.
(handle_one_xevent): Don't check frames; reset on all XI_Enter
and XI_Leave events.
Problem reported by Tino Calancha (Bug#58960).
* src/callproc.c (call_process): Initialize SIGCHLD handling
before possibly creating a child with emacs_span. This need not
be in the critical section that calls emacs_spawn, so do it
outside the critical section.
* src/process.c (child_signal_init): Now extern.
* src/xterm.c (x_probe_xfixes_extension): Rename to
`x_fixes_pointer_blanking_supported', as fixes is now used for
selection tracking as well. All callers changed.
(x_focus_changed): Document where this code is used.
* src/buffer.c (adjust_overlays_for_insert)
(adjust_overlays_for_delete): Repeat for all buffers sharing the same text.
* src/itree.c (itree_insert_gap, itree_delete_gap): Allow an empty tree.
* test/src/buffer-tests.el (buffer-tests--overlays-indirect-bug58928):
New test.
* src/image.c (image_clear_image, lookup_image): Fix coding
style.
(x_destroy_x_image): Remove unnecessary assertion. Call
XDestroyImage, since otherwise only the image data is freed.
(image_from_emacs_colors): Rename variables to make more sense.
Frame GCs are still leaked, but there is nothing that can be
done in that case because the display connection is dead.
* src/xterm.c (x_delete_terminal): Always free xkb desc and
modmap.
This fixes bug #58739. Make subr-arity return, e.g., (12 . 12) rather than
(12 . many) for a function with a fixed number of arguments more than 8.
* lisp/emacs-lisp/comp.el (comp-prepare-args-for-top-level): Only return a cdr
of 'many when there are &rest arguments.
* src/eval.c (eval_sub): Also check for a fixed number of args over 8 when
using the nargs + *args calling convention.
(funcall_subr): Also check numargs <= 8 before using the fixed args calling
convention. Include the case numargs > 8 in the aMany calling convention.
* src/lisp.h (DEFUN): Amend the comment about MANY.
* src/nsterm.m (ns_draw_glyphless_glyph_string_foreground): Work
around silly GCC bug causing splurious warnings.
([EmacsWindow dealloc]): Detach tool bar before dealloc'ing
window. (bug#58857)
The new code to make interpreted closures safe-for-space introduced
a regression in `cconv-tests-interactive-closure-bug51695`, only seen
when using TEST_LOAD_EL.
A few other issues were found and fixed along the way.
* lisp/emacs-lisp/cconv.el (cconv-fv): Change calling convention and
focus on finding the free variables.
(cconv-make-interpreted-closure): New function.
* lisp/loadup.el: Use `compiled-function-p` rather than
`byte-code-function-p` so we also use safe-for-space interpreted
closures when we build with native compilation.
(internal-make-interpreted-closure-function):
Use `cconv-make-interpreted-closure`.
* src/eval.c (syms_of_eval): Rename `internal-filter-closure-env-function`
to `internal-make-interpreted-closure-function`.
(Ffunction): Let that new var build the actual closure.
* test/lisp/emacs-lisp/cconv-tests.el
(cconv-tests-interactive-closure-bug51695): Test specifically the
interpreted case.
Introduce a new 'sqlite-error' and use it for all errors signaled in
'src/sqlite.c', except those that already used 'sqlite-locked-error'.
Include the values of 'sqlite3_errcode', 'sqlite3_extended_errcode',
'sqlite3_errstr' and 'sqlite3_errmsg' in the error data.
* src/sqlite.c (load_dll_functions): Load 'sqlite3_extended_errcode'.
(sqlite-load-extension): Use 'xsignal1' as required by argument type.
(syms_of_sqlite): Introduce a new error type 'sqlite-error'.
(check_sqlite, sqlite-open, bind_values, sqlite-execute)
(sqlite-select, sqlite-load-extension, sqlite-next): Use it.
(sqlite_prepare_errdata): New function.
(sqlite_prepare_errmsg): Remove function.
(sqlite-execute, sqlite-select): Use new function.
(sqlite-locked-error): Derive from 'sqlite-error'.
* src/xterm.c (xi_populate_device_from_info)
(xi_reset_scroll_valuators_for_device_id, xi_handle_device_changed)
(handle_one_xevent):
* src/xterm.h (struct xi_scroll_valuator_t): Get rid of
`pending_enter_reset', which was extremely convoluted and kept
hitting server bugs. Now, valuators are reset upon all crossing
events.
* src/indent.c (scan_for_column, compute_motion): Don't miss
compositions when the position to check for composition happens to
be inside invisible text. (Bug#58791)
* src/xdisp.c (unwind_format_mode_line): Check that
old_top_frame is live before selecting it. It could have been
deleted, which leads to signals when the post-command-hook
installed by xterm-set-window-title runs upon the last frame of
a terminal being deleted.
(gui_consider_frame_title): Rearrange code to be more
understandable.
