* lisp/face-remap.el (face-attrs--make-indirect-safe):
(clone-indirect-buffer-hook): Set up a face remapping alist
(bug#53294).
* lisp/simple.el (clone-indirect-buffer): Move the point the hook
is run.
* src/buffer.c (Fmake_indirect_buffer, syms_of_buffer): Move the
place where the clone-indirect-buffer-hook variable is defined, so
that we can call it from C.
When an Eshell command contains an asynchronous subcommand (such as
calling an external process), it must be evaluated iteratively. See
bug#30725.
* lisp/eshell/esh-cmd.el (eshell-invoke-command): Move most of the
logic from here...
(eshell--invoke-command-directly): ... to here. Also add checks for
subcommands.
* test/lisp/eshell/eshell-tests.el (eshell-test--max-subprocess-time):
New variable.
(eshell-wait-for-subprocess): New function.
(eshell-command-result-p): Use 'eshell-wait-for-subprocess'.
(eshell-test/interp-cmd-external): New test (bug#30725).
* lisp/help-fns.el (help-fns--insert-bindings): New function.
(help-fns--key-bindings): Split menu/key handling and output menu
bindings separately (bug#52870).
* doc/emacs/haiku.texi (Haiku Basics): Update documentation.
* src/haiku_support.cc (Highlight): Apply the hack used for
regular menus to the menu bar as well.
* src/haikumenu.c (digest_menu_items): Set help text on menu bar
if `tooltip-mode' is t and system tooltips are used.
(syms_of_haikumenu): New symbol `tooltip-mode'.
Fchar_syntax did not convert unibyte characters to multibyte when the
current buffer was unibyte, in contrast to `char-syntax` in
byte-compiled code (bug#53260).
* src/bytecode.c (exec_byte_code): Call out to Fchar_syntax;
the dynamic frequency is too low to justify inlining here, and it
did lead to implementations diverging.
* src/syntax.c (Fchar_syntax): Convert non-ASCII unibyte values to
multibyte.
* test/src/syntax-tests.el (syntax-char-syntax): New test.
* doc/misc/eshell.texi (Built-in commands): Expand on the
documentation of echo (bug#27361).
* lisp/eshell/em-basic.el (eshell-echo): Respect OUTPUT-NEWLINE even
when 'eshell-plain-echo-behavior' is non-nil.
(eshell/echo): Add "-N" option and recommend its use over "-n" in
Lisp-friendly echo.
(eshell/printnl): Simplify; 'eshell-stringify' is equivalent to
calling 'eshell-echo' here.
* lisp/menu-bar.el (menu-bar-file-menu): Make the undelete frame
mode option a toggle. (bug#53382)
Also enable some options that were disabled on NS, but should no
longer be since NS now supports the tab bar.
* lisp/international/textsec.el (textsec-domain-suspicious-p):
Ensure that we're not confusing the user if there's a directional
override in the string we're checking.
* lisp/gnus/gnus-art.el (article--check-suspicious-addresses): Be
more resilient to junk in headers. (But all this should be rewritten
eventually, because the entire process is way too convoluted.)
* src/image.c (x_create_xrender_picture):
(Fimage_transforms_p): Use `xrender_supported_p'.
* src/xterm.c (x_term_init): Query for XRender support and
extension versions.
* src/xterm.h (struct x_display_info): New fields
`xrender_supported_p', `xrender_major' and `xrender_minor'.
* lisp/emacs-lisp/bytecomp.el (byte-to-native-output-buffer-file):
Rename from `byte-to-native-output-file'.
(byte-write-target-file): Update.
(byte-compile-file): Write the temporary .elc only when not native
compiling.
(byte-compile-file): Kill the .elc buffer only when not native
compiling.
* lisp/emacs-lisp/comp.el (batch-byte+native-compile): Instead of just
renaming the temporary .elc make use of `byte-write-target-file' to
write it down and kill the temporary buffer.
* lisp/gnus/gnus-art.el (gnus-treat-suspicious-headers): New user
option.
(gnus-article-treat-suspicious-headers): New function.
(article-decode-encoded-words): Hook into the machinery to check
headers.
(article--check-suspicious-addresses): New function.
* lisp/international/textsec-check.el (textsec-propertize): New
function.
(textsec-check): Only check, don't alter STRING.
* lisp/international/textsec.el (textsec-url-suspicious-p): New
function.
* lisp/net/shr.el (shr-tag-a): Mark suspicious links.
* lisp/international/textsec.el (textsec-domain-suspicious-p):
Consider domain names that are whole-script confusables with ASCII
to be suspicious. (I think this is what the Unicode standard is
recommending, but I'm not 100% sure.)
* lisp/international/textsec.el
(textsec-email-address-suspicious-p): Made into its own function.
(textsec-email-suspicious-p): Use it and adjust doc strings.
