`diff-hl-update-async` was enabled in f2696d7, causing a regression
where with-editor buffers wouldn't get cleaned up properly. This would
particularly affect Magit's COMMIT_EDITMSG buffers. To quote jds on
Discord:
It seems like after committing (with cc in the magit buffer), it
leaves COMMIT_EDITMSG around. The next time I try & commit, a single c
keypress immediately jumps to the old COMMIT_EDITMSG buffer, but it's
inactive - C-c C-c closes that buffer but makes no change to git.
I opt for advice instead of find-file or with-editor-mode hooks to
restore normal behavior should with-editor-mode be later disabled
without the death of its buffer (not relevant to magit, specifically,
but it might be to other consumers of with-editor, now or in the
future).
Ref: dgutov/diff-hl#213
Amend: f2696d7302
Wrap doom-customize-theme-hook and use that wrapper also when the local
doom--customize-themes-h-* hook is called immediately, otherwise
settings will be written to custom-file, which is what I've been trying
to avoid with all these takes!
Amend: 0b13525252
Ref: #7929Close: #7933
Co-authored-by: real-or-random <real-or-random@users.noreply.github.com>
Setting custom--inhibit-theme-enable to t isn't enough, since
custom--should-apply-setting will return non-nil if THEME is 'user.
Ref: #7929
Amend: 68771150ba
Monkey-patches counsel-rg to ignore non-zero exit codes. This may have
other side effects, but despite it still producing workable resutls, the
command will discard all its results if there are any errors while
scanning a directory.
Ref: abo-abo/swiper#2339
Fix: #3038
These advice were unnecessary when not only does diff-hl expose
variables to change them, but one of the overridden definitions is
identical to the function it's replacing.
Also includes some other minor refactors.
The NS build of Emacs (on MacOS) seems to struggle rendering our old
fringe bitmaps for diff-hl, leading to noticably slow-downs when
scrolling in buffers with diff-hl enabled. By pre-computing the pixel
height and width of the bitmap, rather than relying on passing `repeat`
to define-fringe-bitmap's ALIGN argument, this seems to spare it a chunk
of wasted cycles.
Ref: https://www.reddit.com/r/emacs/comments/1do1wgj/fringes_incredibly_slow_on_emacs_ns_with_macos/Fix: #7923
Enable `markdown-fontify-code-blocks-natively`, which was disabled in
f5570db due to performance issues, which I try to fix by suppressing
mode hooks on `markdown-fontify-code-block-natively`.
Amend: f5570db4c2Close: #7930
Co-authored-by: real-or-random <real-or-random@users.noreply.github.com>
`custom-theme-set-faces!` calls `custom-theme-set-faces`, which will
mark the customizations for saving to custom.el. But
`custom-theme-set-faces!` is not intended for permanent customizations.
I fix this by setting `custom--inhibit-theme-enable` to prevent some of
the extra work `custom-theme-set-faces` does around changing faces, and
call face-spec-set manually.
Also renames doom--custome-theme-set-face to doom--normalize-face-space,
to better reflect what it does.
Close: #7929
Co-authored-by: real-or-random <real-or-random@users.noreply.github.com>
SPC is equivalent to y when y-or-n-p prompts you, which seems way too
easy to hit by accident, especially with our leader key being (or
involving) SPC by default.
Doom inserts some expensive hooks in MODE-local-vars-hook, like
triggering LSP servers, tree-sitter, or visual enhancements -- things
that are unnecessary in temporary (invisible) buffers, so I suppress
them altogether there.
Toward our eventual goal of moving smartparens out of core, I've adapted
this from code provided by hpfr on Discord, which was adapted from
smartparen's syntax-ppss caching logic. `:config default` will need need
some attention before we can fully move smartparens to its own `:editor
smartparens` module.
Co-authored-by: hpfr <hpfr@users.noreply.github.com>
This one was sneaky. `doom-plist-merge` was mutating the second plist
fed to it, causing issues upwind of its uses. In #7925, for example,
calling `doom-package-recipe` to read a package's recipe would end up
altering it, copying sub-properties of :recipe to other packages'
recipes.
If you've hit #7925, you'll also need to delete your build-cache to get
around the error. I.e.
rm -f $EMACSDIR/.local/straight/build-*-cache.el
doom sync # or upgrade
Fix: #7925
Auctex added a make build step to it's installation. Rather than just
assume the user has the GNU variants of make, sed, and other unix
utilities on their system, I simply generate tex-site.el in elisp,
similar to what we do with org and its org-version.el and
org-loaddefs.el files.
This is a stop-gap measure until I've deployed a mirror+CD on @doomelpa.
change `+format--lsp-fn` to run `+format--lsp-fn` in context of main
buffer rather than apheleia's scratch buffer so that the lsp mode can
be discovered
BREAKING CHANGE: This commit introduces two changes, a breaking, one
not.
The breaking change: `+format-with-lsp` used to control lsp-mode/eglot
integration for this module, but it is now gated behind a new +lsp flag.
Users will need to add it to their `doom!` blocks (in $DOOMDIR/init.el)
to restore the integration.
The other change: I've merged the former `lsp` and `eglot` formatters
into a single `lsp` formatter that dispatches to the appropriate
backend, as well as wrapping this integration in a
`+format-with-lsp-mode` minor mode, so it can be toggled at will; mainly
to assist in debugging formatter behavior.
A unified formatter makes it easier for folks to configure
`+format-with` on a per-project/directory/file basis, without needing to
know what backend this module uses, and opens us up to integrating other
LSP backends in the future (like lsp-bridge).
The former command, magit-unstable-file, would prompt the user to select
a file to unstage, but the intention of this keybind was always to
"unstage *this* file". magit-unstage-buffer-file didn't exist at the
time this keybind was created, so I settled on magit-unstage-file as a
compromise, which is no longer necessary.
I want both `apheleia-formatter` and `apheleia-inhibit` exposed for easy
setting from directory/file-local variables (like .dir-locals.el), but
I'd like to abstract away the implementation (Apheleia) a little, hence
these aliases. This way, whether or not we're still using Apheleia in a
year or three, `+format-with` and `+format-inhibit` should always work.
In commit 5289861109 `+workspace-delete` was
replaced by `+workspace-kill`.
The changes were not complete, though, and mu4e does not start anymore.
This patch completes the changes and should restore mu4e.
Amend: 5289861109
If DIR is not a project and does not have root markers, create a
transient project instead. This ensures that `project-find-file-in`, and
by extension vertico, is still used in non-project directories instead
of falling back to `find-file`
Ref: 1b0af3bfc7
Rather than wrap Apheleia in custom formatting logic, I now use
Apheleia's own machinary to integrate into LSP and Eglot, which is less
complexity to maintain. It also makes settings +format-with a more
reliable option for per-project or per-file configuration.
This also adds a +format/org-src-block command, which I'll incorporate
into the org module in a follow-up commit.
Ref: #7685
