Emacs 28 introduced the much faster native function
buffer-line-statistics, which makes so-long's job significantly less
expensive, so we can afford a larger threshold there.
Still, we gimp it a little if native-comp isn't present.
When selecting a buffer in another workspace with
+vertico/switch-workspace-buffer, that workspace will be switched to,
instead of opening the buffer in the current workspace.
+vertico/switch-workspace-buffer was hardcoded to only list buffers from
the first 9 workspaces. This removes that limit.
Minor catch: workspaces beyond 9 will use lower case a-z as narrowing
keys, followed by upper case A-Z. There will not be any valid narrowing
keys beyond 61 workspaces -- but who in the world would have that many?
Introduced in dcae28c83a, because ignore-errors was being used for the
wrong epg-* function.
Why this happens: epg-make-context throws this error if you don't have a
GnuPG environment set up.
Fix: #6114
Amend: dcae28c83a
dired-virtual-revert (the revert handler for dired-virtual buffers)
simply errors out and prompts the user if they want to switch to a
"real" dired buffer. This isn't helpful the automated nature of
auto-revert, so silence it.
BREAKING CHANGE: The +org-roam-open-buffer-on-find-file variable was
renamed to +org-roam-auto-backlinks-buffer *and* is now disabled by
default. When this is non-nil, it will open the *org-roam* backlinks
side window when roam files are visible, and close it when they aren't.
This change also makes this behavior a little more robust, but is
understandably not everybody's (read: most people's) cup of tea, so it
is now opt-in.
- No longer hard-code fontification of tags and types in roam
completion.
- Prefix types with @ and tags with # -- makes them easier to search for
in completion -- and swap types and hierarchy columns. I.e.
before: TYPE TITLE [TAG [TAG...]]
after: TITLE @TYPE [#TAG [#TAG...]]
- Exclude unwanted (i.e. meta) tags from public display, like ATTACH,
ARCHIVE, or anything in org-num-skip-tags.
This was fixed in Emacs 28.0.91, but the advice will remain for 27
users. At least until we drop support for it.
Fix: #4977
Ref: emacs-mirror/emacs@89d02841ab
Our hacks were too eagerly clearning evil-mc's state. These changes will
ensure that doesn't happen.
Perhaps its time I package this up into a PR upstream?
Fix: #6091
M-x doom/info marks modules that live in $DOOMDIR with &user, but what
if the user has moved their ~/.emacs.d into $DOOMDIR? As this seems to
happen often with Chemacs users, I've made this check a little more
procise.
BREAKING CHANGE: doom-set-jump-maybe-a would formerly set a jump point
if FN returns non-nil. That condition was changed to: if the current
buffer or cursor position have changed after FN.
Doom sets jump points before following most org links. This allows you
to jump back to where you were afterwards using
C-o (better-jumper-jump-backward) or forward with
C-i (better-jumper-jump-forward). It does this by setting a jump point
on the org-open-at-point-functions hook.
However, not all org links' :follow handlers trigger this hook (like id:
or roam: links), and you can get around it by triggering link opener
handlers directly (with M-x org-goto). Most of these link openers use
org-mark-ring-push to record jump points in org's own jumplist, so
advising it to call better-jumper-set-jump should address more of these
fringe cases.
Fix: #6098
