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
To be more consistent with other uses of the verb in the Emacs
ecosystem. Also done in preparation for a new +workspaces/delete command
for #7869.
Ref: #7869
BREAKING CHANGE: This removes the lsp-python-ms package (along with it,
lsp-python-ms support), since the package no longer works in recent
versions of Python and is no longer maintained upstream.
BREAKING CHANGE: This commit removes the ob-ipython package; it's been
deprecated for some time because the upstream package is no longer
maintained and is broken on most, common versions of Python anyway.
Jupyter is a superior alternative, in any case.
This removes the php-cs-fixer.el package. This is not a breaking change
because the package hasn't been this module's default formatter since
4ecd616. It has since been replaced by phpcbf (provided by
squizlabs/php_codesniffer).
Amend: 4ecd616cd8
This module has been deprecated for some time because it is too
trivially small, and it's always been unclear how hydra fits into Doom's
vision for its UI, but I intend to lean into either transient or Embark
in the future.
BREAKING CHANGE: This removes the org-yt package, which will break
[[yt:...]] links in org, if you're using them. The package is a source
of errors, likely because it hasn't been updated for Org's new element
parsing system.
This package was originally included for eventual use in Doom's
documentation, but it likely won't be used and isn't universally useful
enough to keep in this module by default.
Fix: #7913Fix: #7317
BREAKING CHANGE: This removes the pcre2el package, which Doom was using
solely for one function to escape PCREs. In the interest of thinning out
Doom's core, I've hoisted a simpler version of the function into Doom's
stdlib so I can remove the dependency.
Folks seem to expect to *only* see the current project in the treemacs
pane when they open it. I sort of agree, so I've tweaked
+treemacs/toggle (on SPC o p) to use
`treemacs-add-and-display-current-project-exclusively` instead.
Fix: #7614
This module has been deprecated for some time, because the module is
much too trivial to warrant being its own module. At most, it may find a
new home in doomemacs/modules-contrib, but for now, it doesn't belong in
the official module library.
This module has been deprecated for some time; it's too niche, too
trivial, and is mostly centered around a single hydra (official support
for which I'll be phasing out of Doom).
Previously, `+lisp/find-file-in-quicklisp` used a hard-coded directory
aimed at a nonstandard Quicklisp install location, `~/.quicklisp`. The
default install location is `~/quicklisp`, without the dot. This commit
makes the user's Quicklisp directory configurable through the variable
`+lisp-quicklisp-paths`. `+lisp/find-file-in-quicklisp` has been updated
to respect this variable.
Close: #7830
Co-authored-by: bpseudopod <bpseudopod@users.noreply.github.com>
I'm considering removing anaconda entirely from the module (to lean on
LSP), but for the time being, this should help with anaconda taking over
ine eglot-enabled buffers.
Fix: #7823
This change (made in ae9e1fea) was a tad too hasty, since
general.el (under the hood) implicitly deferred the keybinding until
eglot-mode-map, which define-key does not do.
Fix: #7911
Amend: ae9e1feaa6
Without this change, we need to pass a prefix arg that is one more than
the actual fold-level we're trying to achieve; `C-u 2` is equivalent to
not passing a digit arg. While this looks intentional, I don't see what
purpose it serves.
This change provides saner behavior that matches other commands that
handle fold-levels (eg. `org-fold-hide-sublevels`).
With four folding systems in play, there are numerous ways to create
nested folds (eg. hideshow folds inside folded outline headings with
`+fold/close-all`). So `+fold/open-rec' needs to be able to open all
kinds of folds within the newly unfolded region. It can't stop at
unfolding the type of fold it was called on.
This module is also used by non-Evil users. Also,
`evil-vimish-fold/next-fold` and friends just wrap the corresponding
vimish-fold functions in `dotimes`, anyway.
