setenv treats the TZ (and only TZ) envvar especially, so we have to too,
since I'm intentionally avoiding iteratively setenv'ing envvars to avoid
the unnecessary extra work it does.
Fix: #5760
Reduce cases where native-comp procedures were executed in
non-native-comp contexts (somehow).
Also:
- Adds a NATIVECOMP constant for statically detecting the feature.
- Remove native-comp-* -> comp-* aliases (Emacs HEAD has long since
moved on from these).
- Add 'no-native-compile: t' to autoloads file.
The message simply test for the "not-a-version-yet" patch number and
displays an explanation to prepare users for raised error frequency and
having a backup version of Emacs if they do not want to deal with
issues.
doom--fix-broken-smie-modes-a forces dtrt-indent-mode to expect exactly
one argument so an error will be raised when the latter is turned on
without args programmatically in lisp.
This generalizes and cleans up the linter API so that it can be used in
other Doom projects (or CLI commands, like our WIP changelog generator).
Besides that, our git's commit conventions saw two changes:
- A new 'Amend' keyword, for indicating a commit corrects an earlier,
recent one. This should be used to omit (or merge) commits in the eyes
of the changelog generator.
- Trailers must now follow the 'KEY: VALUE' format, which is supported
OOTB by 'git interpret-trailers' and makes scraping them much eacher.
Before, omitting the colon was mandatory, this is no longer the case.
Other highlights:
- For linter rules: replaced :footer and :refs keys with :trailers (a
string->string alist). Invalid trailers will be left in BODY's tail.
- Added a linter for colon delimiters in commit trailers (along with
other formatting checks, like capitalization and one-per-line checks).
While lists are technically cons cells, cons cells don't have all the
properties of lists, so doom-enlist shouldn't treat it as one.
Before:
(doom-enlist '(a . b)) #=> (a . b)
After:
(doom-enlist '(a . b)) #=> ((a . b))
Before this fix:
(fn! (x &key y z))
;; implies
(fn! (&key x &allow-other-keys)).
But
(fn! (x (&key y) &key z))
;; would not imply
(fn! (x (&key y &allow-other-keys) &key z &allow-other-keys)).
68d8364ae added a deprecation notice to any use of a
MODE-local-vars-hook, instructing users to use after-MODE-hook instead,
however this hasn't been decided yet and shouldn't have made it into
68d8364ae.
Ref 68d8364aeaFix#5612
Fixes a regression introduced in 0bb4d4dfc, which changed how
doom-switch-buffer-hook works, but makes it fire too early during
startup, eagerly loading packages that wait for the first interactively
opened buffer. This moves doom-init-ui-h to
window-buffer-change-functions, which fires once at the end of startup,
when any buffer is displayed for the first time in the session.
Ref 0bb4d4dfcb
Emacs 27 introduced a bunch of `window-*-change-functions` hooks,
including `window-selection-change-functions` and
`window-buffer-change-functions`, which handles 98% of the use case for
Doom's `doom-switch-{buffer,window,frame}-hook` hooks, so I've rewritten
them to use them under the hood, which amounts to simpler code and fewer
hacks.
Use the new server-after-make-frame-hook, introduced in Emacs 27.1, to
clean up how we initialize fonts, themes, and the UI in general.
This may indirectly fix font issues in daemon sessions.
Doom doesn't use these two functions, but since they can be useful for
debugging *and* straight provides impleemntations for them already,
they're trivial to fix.
