The method for customizing +notmuch-sync-backend was changed in 19d4126
to accept a string; consequently the current example in the docs results
in an error:
user-error: Invalid notmuch backend specified: custom
Ref: 19d41262e4
In Emacs 28+, the mode-name in emacs-lisp-mode is "ELisp/X" (where X = d
or l depending on lexical-binding). I find this much more useful than
"Emacs-Lisp" in <=27.x or our static replacement "Elisp".
Permit `;;* ...` be recognized by imenu and outline-minor-mode (and
outline's commands). This also patches Lispy to reflect this new
configuration (if :lang emacs-lisp is active).
Close: #6732
Co-authored-by: LemonBreezes <LemonBreezes@users.noreply.github.com>
This allows for more rapid 'doom sync'ing when testing different module
combinations, but puts the onus on the user to run `doom sync
--purge` (or `doom purge`, which will later be renamed `doom gc`) to
prune orphaned packages/repos.
Due to some packages and modules using this in some (non-GUI) builds of
Emacs. display-graphic-p isn't enough here, so rather than police all
the possible offenders, I've defined it to no-op in those cases,
instead.
Close: #6876
Ensure that highlight-indent-guides-auto-set-faces is called when the
active theme and frame has been initialized (particularly tricky in
daemon sessions). With this assurance, we don't need to suppress its
errors anymore.
Close: #6900
Co-authored-by: Sleepful <Sleepful@users.noreply.github.com>
use-package evaluates :preface *before* :when, so the two lsp! hooks
would run unconditionally, causing errors and warnings when opening a
haskell-mode buffer.
*
modules/lang/haskell/config.el (haskell-literate-mode-local-vars-hook):
add lsp! to hook.
Leave it to modeline modules, plugins, and/or the user to support this
plugin, rather than impose winum's (rather destructive) side-effects on
them all.
This regression was introduced in 5a5195b; it broke `package!`s ability
to override the recipes of packages declared by other modules, due to
some API assumptions that aren't true yet (in unpushed post 3.1 work).
Amend: 5a5195b84dFix: #6901
The add-node-modules-path package calls `npm bin` to locate the
node_modules/.bin, and does so while invoking the user's $SHELL, which
can be very expensive depending on the user's shell configuration,
possibly adding seconds to the startup time of any JS/TS file.
To mitigate this, I replace the package with a much faster, and in-house
heuristic. Folks with more complex needs should be using direnv anyway.
Fix: #6878
Sync definition of __DOOMGEOM in bin/doomscript with corresponding
variable of bin/doom.
* bin/doomscript (__DOOMGEOM): split tput call into two separate calls
Ref: beef0aef02
The current config is kinda confusing, it uses the same symbol for both
forwarded/passed mails and replied mails. FontAwesome provides a nice
icon for "reply", why not use it!
For some reason, the transient-append-suffix adding magit-worktree back
to magit-dispatch — after its potential replacement by
magit-gitflow-popup — was having no effect. (It does when moved into
(after! magit-gitflow), so the issue must have something to do with when
transient-append-suffix is called.) magit-worktree wasn't appearing in
the magit-dispatch popup when magit-gitflow was enabled, nor was the '*'
keybind for magit-worktree in effect outside (or inside) the popup,
unlike '%' for magit-gitflow-popup.
Replace the ineffectual transient-append-suffix with a normal and visual
mode keybind for magit-worktree in magit-mode-map (and move the
unconditionally defined keybind for magit-gitflow-popup into (after!
magit-gitflow)). Also, append the magit-gitflow-popup transient suffix
to magit-worktree instead of replacing it, so that the latter still
appears in magit-dispatch (though under the original keybind 'Z' — which
isn't really an issue, since evil-collection-magit doesn't seem to
update the keybinds of any of the other commands accessible from
magit-dispatch — but also callable with '*').
warning-suppress-types' pre-29 documentation suggests that it accepts a
list of symbols, but a recent, upstream correction changes this.
Ref: emacs-mirror/emacs@d5ee49c25c
Amend: 8c442d84b9
C-S-s while company is completing shoudl bring up the results in your
completion framework of choice (ivy, helm, vertico, etc), but failed to
do so for vertico (for any completion backend besides company-capf
perhaps).
BREAKING CHANGE: Removes the eval-when-compile! macro. I recently
discovered cl-eval-when and it serves this macro's purpose well
enough (and is autoloaded+built-in).
This macro wasn't used anywhere in (published) Doom code, however, and
users shouldn't really be using it, so its removal is unlikely to cause
any actual breakage.
a4b5831 removed this let-binding because I expected the function to
capture it, which seems to be the case in general, except for #6865.
Since add-transient-hook! could conceivably be used in an elisp file
with lexical-binding off, I'll take fewer chances and live with a little
redundancy.
Amend: a4b58311daFix: #6865
