With current hacks involving `org-agenda-files`, one cannot use org agenda
commands like clock in or change the entry's state to done, etc.
Specifically, the mangled behavior observed at #4901. This commit fixes those
problems around `org-agenda` buffer specific commands.
For more details, please refer to https://github.com/hlissner/doom-emacs/issues/4759#issuecomment-822100632.
When opening an org file and the point is placed in a folded region (by
saveplace), it reveals the current subtree. Instead, it will reveal the
current subtree *and* its ancestors (and their siblings), which is more
intuitive.
This could potentially cause stuttering for long lived sessions where
the user takes few breaks, but could yield great performance benefits in
the shorter term. We'll have to experiment.
I was conservative before. I didn't want to change this globally in case
it had a negative performance impact on shy servers that don't talk
much, but this turned out to be too paranoid. Untalkative servers are
barely, if at all, affected, and chatty ones perform better across the
board.
I've noticed that some docs are a bit outdated, as homebrew doesn't have
`brew cask` command now, all features regarding to casks were moved to
the option `--cask` instead.
This commit just introduce such changes to be up-to-date with homebrew instructions.
Some keys were bound to no-longer-existing commands, or commands which changed
name.
Others only made sense in the REPL, or /also/ in the REPL, so the keybinds are
now in `scheme-mode-map', `geiser-repl-mode-map', or both accordingly.
Some modes will set comment-line-break-function to an anonymous function
or advice (as octave-mode does), which will cause this fboundp check to
throw a type error when you press RET while on a commented line.
Due to case shenanigans, ob-C.el's feature name is ob-C, not ob-c,
preventing folks from using 'c' (lowercase) as a language name for org
babel blocks.
- otherwise, `+lookup-online-backend-fn` would be called as well
- lsp-describe-thing-at-point was marked as not async in
d4eb7e31ac
- but `lsp-describe-thing-at-point` calls `lsp--make-request`
internally (which is async I assume):
eda51c2166/lsp-mode.el (L4677)
emacs-geiser/geiser@aa26163 -> emacs-geiser/geiser@c5a9aae
With emacs-geiser/geiser@c5a9aae, geiser-impl--add-to-alist is now
autoloaded. This fixes#4863 by preventing void-function errors at
startup or 'doom sync'-time.
But this presents a new problem: geiser-impl is guaranteed to be
needlessly loaded at startup, the first time geiser-impl--add-to-list is
called among, say, geiser-chez's package autoloads. To get around this I
forcibly redefine it as an inline-able function (with defsubst) and
shove this new definition into Doom's autoloads file (where it will
appear before package autoloads).
Since Doom byte-compiles its autoloads file (containing concatenated
autoloads from all doom modules and packages), geiser-impl--add-to-alist
calls (and geiser-activate-implementation calls too, which is already
inline-able) will be inlined, fixing the issue.
Fixes#4863
+ Adds electric alignment of amounts on RET
+ Fixes all of beancount's config (which was running because the package
is called beancount, not beancount-mode)
+ Fixes flood of "invalid face org-level-N" warnings
+ Fixes beancount-electric-currency for appending currency to unlabelled
amounts. Also enables this to fall back to first operating_currency in
document, in available
+ Binds TAB to beancount-outline-cycle on outline heading, otherwise
indent-according-to-mode. No need for completion popup when company
already provides this.
