* src/print.c (print_vectorlike): Make sure module function's
address prints with a leading "0x". This fixes emacs-module-tests
on MS-Windows. Fix whitespace.
* src/dynlib.c (dynlib_addr): Remove unused variable. Update
commentary.
Reuse existing functionality and simplify the code a bit.
* src/emacs-module.c (Fmodule_load): Use unwind protection to clean up
runtime object.
(funcall_module): Use unwind protection to clean up environment
object.
(finalize_environment): Simplify signature.
(finalize_environment_unwind, finalize_runtime_unwind): New functions.
Most of these seem to run afoul of the comment "Do NOT use
'eassert' for checking validity of user code in the module."
* src/emacs-module.c (MODULE_FUNCTION_BEGIN_NO_CATCH)
(module_non_local_exit_check, module_non_local_exit_clear)
(module_non_local_exit_get, module_non_local_exit_signal)
(module_non_local_exit_throw, module_make_string):
Remove unnecessary easserts that pointers are nonnull.
Hardware checks this for us nowadays, and the checks
just clutter up the code.
(module_extract_integer): Remove unnecessary verify that
a C signed integer is in the range INTMAX_MIN..INTMAX_MAX.
The C standard guarantees this.
(module_copy_string_contents): Remove unnecessary eassert
that Lisp strings are null-terminated.
(module_function_arity): Remove unnecessary easserts that
function arities are in range.
* src/emacs-module.c (module_copy_string_contents):
Remove checking, as string lengths are always nonnegative and less
than STRING_BYTES_BOUND, and this is checked elsewhere.
(module_make_string): Check length against STRING_BYTES_BOUND, a
tighter bound than MOST_POSITIVE_FIXNUM. (funcall_module): Don't
assume that an out-of-range integer is nonnegative.
Tell the compiler that SCHARS and STRING_BYTES are nonnegative, in
the hopes that this will optimize a bit better. Also, check this
at runtime if ENABLE_CHECKING.
* src/lisp.h (SCHARS, STRING_BYTES):
eassume that these functions return nonnegative values.
(STRING_SET_CHARS) [ENABLE_CHECKING]:
eassert that newsize is nonnegative.
The idea is that modules should call env->should_quit from time to
time and return as quickly as possible if it returns true.
* src/emacs-module.c (module_should_quit): New module function.
(initialize_environment): Use it.
(funcall_module): Process potential pending quit.
* src/eval.c (maybe_quit): Add reference to module_should_quit.
In particular, alias violations are likely for the return values of
dlsym(3), which get cast around arbitrarily.
* src/emacs-module.c (Fmodule_load): Use ATTRIBUTE_MAY_ALIAS.
* src/emacs-module.c (Fmodule_load, funcall_module): Adapt callers.
(finalize_environment): Add parameter for public part of the
environment, like 'initialize_environment'. Add assertions.
Fix a FIXME in emacs-module.c. Put the printing into print.c, like
other types.
* src/print.c (print_vectorlike): Add code to print module functions.
* src/emacs-module.c (funcall_module): Stop calling
'module_format_fun_env'. Now that module functions are first-class
objects, they can be added to signal data directly.
(module_handle_signal): Remove now-unused function
'module_format_fun_env'.
* test/src/emacs-module-tests.el (mod-test-sum-test): Adapt unit test.
* src/eval.c (funcall_lambda): Adapt call to changed signature of
'funcall_module'.
* src/emacs-module.c (MODULE_FUNCTION_BEGIN_NO_CATCH): New helper
macro.
(MODULE_FUNCTION_BEGIN, module_type_of, module_is_not_nil, module_eq):
Use it.
* doc/lispref/strings.texi (Formatting Strings):
* src/editfns.c (Fformat): Format field numbers no longer need
to be unique, reverting the previous doc change since that has
now been fixed. Also, document that %% should not have modifiers.
* src/editfns.c (styled_format): Improve performance. Remove
the need for the new prepass over the format string, by using
a typically-more-generous bound for the info array size.
Initialize the info array lazily. Move string inspection to
the same area to help caching. Avoid the need for a
converted_to_string bitfield by using EQ. Cache arg in a
local and avoid some potential aliasing issues to help the
compiler. Info array is now 0-origin, not 1-origin.
Previously styled_format overwrite the argument vector. This is no
longer possible because there might be more than one specification per
argument. Use the existing auxiliary info array instead.
* src/editfns.c (styled_format): Record arguments in the info
structure instead of overwriting them.
* test/src/editfns-tests.el (format-with-field): Add unit test.
* doc/lispref/strings.texi (Formatting Strings):
Don’t allow mixing numbered with unnumbered format specs.
* src/editfns.c (styled_format): Don’t bother checking for field 0,
since it doesn’t crash and the behavior is not specified.
* test/src/editfns-tests.el (format-with-field): Adjust tests to
match current doc. Add more tests for out-of-range fields.
This made (string-to-number "10") 20% faster on my old desktop,
an AMD Phenom II X4 910e running Fedora 25 x86-64.
* admin/merge-gnulib (GNULIB_MODULES): Remove strtoumax.
* lib/gnulib.mk.in, m4/gnulib-comp.m4: Regenerate.
* lib/strtoul.c, lib/strtoull.c, lib/strtoumax.c, m4/strtoull.m4:
* m4/strtoumax.m4: Remove.
* src/editfns.c (str2num): New function.
(styled_format): Use it instead of strtoumax. Use ptrdiff_t
instead of uintmax_t. Check for integer overflow.
* src/lread.c (LEAD_INT, DOT_CHAR, TRAIL_INT, E_EXP):
Move to private scope and make them enums.
(string_to_number): Compute integer value directly during
first pass instead of revisiting it with strtoumax later.
* src/editfns.c (styled_format): Allow field numbers in a %% spec.
No need for a special diagnostic for field numbers greater than
PTRDIFF_MAX. Reword diagnostic for field 0.
* test/src/editfns-tests.el (format-with-field): Adjust to match.
A field number explicitly specifies the argument to be formatted.
This is especially important for potential localization work, since
grammars of various languages dictate different word orders.
* src/editfns.c (Fformat): Update documentation.
(styled_format): Implement field numbers.
* doc/lispref/strings.texi (Formatting Strings): Document field numbers.
* lisp/emacs-lisp/bytecomp.el (byte-compile-format-warn): Adapt.
* test/src/editfns-tests.el (format-with-field): New unit test.
* src/xdisp.c (get_next_display_element): Dispatch used format string
for unprintables based on new display-raw-bytes-as-hex variable.
(display-raw-bytes-as-hex): New variable. (Bug#27122)
* lisp/cus-start.el: Add defcustom form for display-raw-bytes-as-hex.
* doc/emacs/display.texi: Document the new variable.
* etc/NEWS: Mention display-raw-bytes-as-hex.
* test/manual/redisplay-testsuite.el (test-redisplay-5-toggle)
(test-redisplay-5): New tests.
(test-redisplay): Call test-redisplay-5.
* src/xdisp.c (get_next_display_element): Dispatch used format string
for unprintables based on new display-raw-bytes-as-hex variable.
(display-raw-bytes-as-hex): New variable. (Bug#27122)
* lisp/cus-start.el: Add defcustom form for display-raw-bytes-as-hex.
* doc/emacs/display.texi: Document the new variable.
* etc/NEWS: Mention display-raw-bytes-as-hex.
* test/manual/redisplay-testsuite.el (test-redisplay-5-toggle)
(test-redisplay-5): New tests.
(test-redisplay): Call test-redisplay-5.
* doc/emacs/display.texi (Horizontal Scrolling): Document the new
mode of auto-hscrolling only the current line.
* src/xdisp.c (init_iterator): When hscrolling only the
current line, apply the window's min_hscroll here, so that
non-current lines will be hscrolled by that minimum.
Suggested by Stephen Berman <stephen.berman@gmx.net>.
(hscroll_window_tree): Account for window's min_hscroll when
deciding whether to recompute the hscroll.
(display_line): Subtract window's min_hscroll from x_incr, as that
was already accounted for in init_iterator. (Bug#27008)
* src/.gdbinit (to_string): Use an unsigned representation for
Lisp values, as requested by Eli Zaretskii (Bug#27098).
Also, use "make_number(N)" for Lisp integers.
Use IN_MASK_ADD instead of using a no-longer-promiscuous-enough
mask. This simplifies the code and restores the ability to
use IN_ACCESS, IN_CLOSE_WRITE, IN_CLOSE_NOWRITE, and IN_OPEN
in some cases (Bug#26973).
* src/inotify.c (INOTIFY_DEFAULT_MASK): Remove.
(Finotify_add_watch): Use IN_MASK_ADD instead.
There was no need to remove it in the 2017-03-26 inotify change,
as it is like IN_DONT_FOLLOW and does not affect other watchers
for the same file.
* src/inotify.c (symbol_to_inotifymask, Finotify_add_watch)
(syms_of_inotify): Bring back onlydir.
