Instead of blindly assuming that all Emacs strings are valid UTF-8,
which they are not, use a more careful conversion going via UTF-16
which is what NSString uses internally. Unpaired surrogates will
still go through to the NSString objects, but the NS libs handle them
gracefully.
* src/nsterm.h (EmacsString): New category.
* src/nsfns.m (all_nonzero_ascii): New helper function.
([NSString stringWithLispString:]): New method.
(ns_set_name_internal): Use new conversion method.
* lisp/vc/diff-mode.el (diff-goto-source): Don't output a status
about the hunk just when jumping to it (bug#38370). This would
output "Hunk already applied" when browsing diffs.
* lisp/time.el (world-clock-mode): Set 'revert-buffer-function'
buffer-locally rather than globally.
(display-time-world): Unobsolete alias for 'world-clock'. Some users
might be used to the old name.
Add a customizable variable for driver options (such as linker flags)
to pass to libgccjit (Bug #42761).
* lisp/emacs-lisp/comp.el (comp-native-driver-options): New
customization variable.
* src/comp.c: Use comp-native-driver-options to set libgccjit's driver
options, if supported on the library's ABI version.
'dlopen' can return the same handle if two shared with the same
filename are loaded in two different times (even if the first was
deleted!). To prevent this scenario the last modification time of the
source file is included in the hashing algorithm.
* src/comp.c (Fcomp_el_to_eln_filename): Update hashing algo to
include the source last modification date.
* src/lread.c (maybe_swap_for_eln): Do not check for eln newer
then elc as this is now unnecessary.
* lisp/simple.el (read-extended-command): Allow doing interactive
searches over the completions (bug#12490). This restores the
behaviour from Emacs 23 that was lost in Emacs 24.
This bug affected compilation of
(cond ((member '(some list) variable) ...) ...)
While equal is symmetric, member is not; in the latter case the
arguments must be a variable and a constant list, in that order.
Reported by Ikumi Keita.
* lisp/emacs-lisp/bytecomp.el (byte-compile--cond-switch-prefix):
Don't treat equality and member predicates in the same way; only
the former are symmetric in their arguments.
* test/lisp/emacs-lisp/bytecomp-tests.el
(byte-opt-testsuite-arith-data): Add test cases.
* test/lisp/emacs-lisp/bytecomp-tests.el (bytecomp-check-1)
(test-byte-opt-arithmetic, bytecomp-lexbind-check-1)
(bytecomp-lexbind-explain-1):
When comparing interpreted with compiled results, don't consider all
errors to be equal; take the error type into account. (The error
arguments may differ, but there may be good reasons for that.)
* test/lisp/emacs-lisp/bytecomp-tests.el
(test-byte-comp-macro-expand-lexical-override): Remove functions
before testing so that the test can be run twice without failing.
* lisp/net/tramp-sh.el (tramp-multi-hop-p, tramp-compute-multi-hops):
Move them from here ...
* lisp/net/tramp.el (tramp-multi-hop-p, tramp-compute-multi-hops): ... here.
(tramp-direct-async-process-p): Use `tramp-compute-multi-hops'.
* lisp/comint.el (comint-highlight-input): New variable (bug#32344).
(comint-send-input): Use it.
* lisp/progmodes/python.el (inferior-python-mode): Set it.
* src/minibuf.c (read_minibuf_noninteractive): Signal an
`end-of-file' error when reading from stdin instead of a general
error (bug#34123). This makes it easier to write code that
recovers from this situation.
Suggested by Noam Postavsky <npostavs@gmail.com>.
* lisp/ido.el (ido-everywhere): Turn on ido-mode, if it's not already
on. Otherwise, having ido-everywhere enabled messes all file and
buffer reading interactive commands (bug#34292).
* lisp/emulation/viper-util.el (viper-check-minibuffer-overlay):
* lisp/emulation/viper-cmd.el (viper-minibuffer-standard-hook)
(viper-minibuffer-real-start, viper-submit-report): No need to
check whether minibuffer-prompt-end is defined.
