The doom-*.el files will soon be moved to separate core modules (whose
load order the user can dictate); when this happens, this block could be
re-evaluated anytime the user calls `doom/reload` or otherwise reloads
the active profile.
Instead of advising load-theme, which was more opinionated, and broke in
instances where users (relying on `custom-enabled-themes`s setter) or
packages (like auto-dark) were enabling themes with `enable-theme`
instead of `load-theme`.
This also adds support for enabling multiple themes with `doom-theme`.
Fix: #8119
Cast too wide a net; there are likely too many *rc files that shouldn't
be treated as shell scripts.
Besides, there's already a fallback *rc rule for conf-mode in
lisp/doom-editor.el.
Revert: bdc35faff2
Assuming GPG is present and set up by default can be surprising for
beginners, so this commit makes GPG integration opt-in, behind a new
+gnupg flag in the :config default module. There'll be more added to
this later.
Regardless of DEEP?, a record's fields should be copied one level deep
at least, to ensure shallow changes to shallow copies don't affect the
original.
Amend: 169540ad3b
Emacs requires '-*- lexical-binding: t -*-' to be prefixed with an elisp
line comment (two semicolons), but due to our shebang shenanigans (which
are, unfortunately, necessary) starting this line with a :;, Emacs
ignores the `lexical-binding: t` at the top, thereby executing bin/doom
without lexical binding.
Until a better workaround is found, our CLI framework will have to
adapt. Fortunately, this only affects bin/doom and not any arbitrary
doomscript.
Error occurs on Emacs 30.0.92 or newer, because this variable (what it's
aliased to, at least) was moved to comp-run.el, upstream.
Ref: emacsmirror/emacs@e6a955d242
cli/gc.el accidentally snuck into 5e84709, along with a particular
v3-isms that doesn't belong in Doom yet, such as autoloads in CLI libs
and them providing subfeatures of doom-cli, rather than doom-cli-*.
These have been reversed for now.
Amend: 5e84709577
`letenv!` is a layover from the days before
`with-environment-variables` (introduced in 28.x), and it remained
afterwards because I preferred the shorter name. From v3 and onward,
Doom's core will be put on a diet which, among other things, will
include culling redundant or superfluous functions/macros like this one.
Many users seem to treat `doom/reload` like a 1-to-1 replacement for
`doom sync` + restart Emacs (which it is not), and are surprised when it
fails to properly "reload" their config. I'm considering removing
`doom/reload` altogether for this reason, or turning it into a light
wrapper around `restart-emacs`, but for now I'll settle for recommending
it less in documentation and comments. The references in docs/*.org will
be dealt with later.
The performance benefit of doing so has always been questionable or, at
best, negligible, but has caused numerous issues over the years. The
latest one being #8162, where byte-compiling a profile init file with
too many package autoloads would consume more than 255 opcodes, causing
an overflow error.
For simplicity's sake, Doom will no longer byte-compile this file.
Fix: #8162
Freeing this up to use for a future command (doom init), and because
'doom install' ought to be used very rarely and deliberately, so it
shouldn't have a convenient alias anyway.
These commands were removed in an ongoing effort to slim down Doom and
its core. The `doom/goto-private-*-file` family of commands were
redundant with `doom/open-private-config` and
`doom/find-file-in-private-config`.
This is an old issue that's haunted Doom for a while. I had initially
planned to wait until the switch to Elpaca, but I decided to just sit
down and solve this.
This ensures package autoloads are always written in depth-first
dependency order to Doom's profile init file, preventing load-order
issues like the notorious void-function geiser-activate-implementation
error. `geiser` needs to be built before any `geiser-*` plugins, since
its plugins reference variables/functions in geiser's own autoloads, but
there's no way to enforce package order in `straight--build-cache`
currently, and subsequent package updates (or just deleting package
directories by hand) can change the order of straight's build-cache in
subtle ways.
Fix: #7693Fix: #7472
The lexical bindings for the load suffix variables in early-init.el does
this work already, so affecting the global state of these variables is
redundant and overkill.
It seems either Emacs' warning library and/or debugger relies on
custom.el functionality at load time to properly function. If
`custom-dont-initialize` is non-nil when a warning or error occurs,
Emacs fails to fully load the `warning` or `backtrace` libraries,
causing this error to obscure the true warning/error with:
Error in delayed-warnings-hook (display-delayed-warnings): (void-variable
warning-minimum-log-level)
`doom-module-load-path` will be removed in the next big refactor commit,
so as a stop gap, I simply move it to doom.el to resolve the reference
error.
This also removes the obsolete alias `doom-modules-dirs`, since it's
been deprecated long enough (and `doom-module-load-path` itself won't be
around much longer anyway).
Ref: #8147
Amend: 8cafbe4408
The issue required a more systemic fix, because the former (before
9e6c46a even) only recorded `Info-directory-list` for the packages that
were installed/updated in that `doom sync` session, forgetting all
packages installed in the past.
Fix: #8143
Amend: 9e6c46a332
Autoloads were accidentally reversed in 114f996. For some packages,
autoload order is signifcant (such as json-mode, which autoloads
json-mode-auto-mode-list first, then modifies it in a later autoload).
Fix: #8143
Amend: 114f99688c
