| Age | Commit message (Collapse) | Author |
|---|
|
Take advantage of recent (Emacs 28) additions to project.el to make a
convenient "magit-to-project" command. Some C-c p bindings are now
redundant with the new C-x p keymap; keep them for a bit, as Emacs 28
is not coming out anytime soon.
Add an ibuffer group for "messaging" buffers.
Set lua indentation to 2 spaces; the Lua Style Guide[1] says that this
convention is used in many learning resources; this is also what
pandoc filters use AFAICT.
[1]: http://lua-users.org/wiki/LuaStyleGuide
Org 9.4 introduces a new face for DONE headlines; I think the DONE
keyword face is enough to distinguish these headlines, so I'll turn
off that feature.
|
|
Turns out somebody filed an issue soon after I committed my
workaround:
https://github.com/dajva/rg.el/issues/72
Now fixed 🙌
|
|
Cf. <https://github.com/dgutov/diff-hl/commit/a625033fb1dde83f6e4c2fc21f632b22ec34b609>.
|
|
|
|
Use C-x o with repeatable o instead. This removes a conflict with
Magit and Org bindings.
Inspiration: https://emacs.stackexchange.com/a/52203/10209
|
|
* Use Custom to set whitespace style.
As much as I dislike using Custom for lists where I just want to
add/remove a couple of elements, the previous method introduced a
subtle bug where visiting a diff buffer, *then* activating
whitespace-mode resulted in diff-mode's whitespace style applying
instead of mine. Something to do with diff-setup-whitespace I guess.
* Make Org comply with RET/C-j conventions in the rest of Emacs.
Since Emacs 24.4, by virtue of electric-indent-mode being enabled by
default, RET is the "smart newline" key which handles indentation, and
C-j is the "dumb newline" key which simply inputs character 0x0A.
* Stop popping another frame for Ediff's control buffer.
The popup frame needs its own .Xresources entry. Sometimes it
disappears behind the system tray. Having another window allows
other-windowing.
* Make line-number-mode try harder with long lines.
* Change the scroll logic.
scroll-conservatively > 100 tells Emacs to *never ever* recenter
point, which is annoying when jumping to e.g. a search result or a
tags definition.
scroll-conservatively = n < 100 tells Emacs to recenter when scrolling
needs to move more than n lines, and to scroll only m lines when m≤n.
* And some other stuff.
|
|
Emacs attempts to expand macros when loading uncompiled files to speed
up their execution. Since rx is not loaded during startup, Emacs is
not aware of the rx pattern for pcase, and takes it for the rx macro.
Since this macro does not understand (let …) forms, it throws an
"unknown rx pattern" error, which interrupts macro-expansion and shows
a noisy warning.
|
|
|
|
|
|
|
|
For reasons not entirely clear, the password prompt would get stuck
when trying to sudoedit a file. minibuffer-depth-indicate-mode told
me that I was 16 prompts deep. I'll chalk if off to my/project-name
calling functions it shouldn't call when TRAMP is busy.
|
|
|
|
As well as a function to generate them.
|
|
I'm about to add another my/make-… function, and it will need to be
defined before those global-set-key calls.
|
|
diff-hl does the Git gutter thing; I don't use hideshow often enough
to care about it being visible in the fringe.
|
|
|
|
Color fonts improve emoji readability IMO. Noto's emoji font does not
cover 🗚 nor 🗛 though.
While in there, use simplify a version check.
|
|
Add additional stuff manually.
|
|
Cute fluff, yet fluff still.
|
|
The code below works with the new indicator, but it's a bit ugly.
(let ((props (text-properties-at
0 (car (alist-get 'compilation-in-progress mode-line-modes)))))
(setf (car (alist-get 'compilation-in-progress mode-line-modes))
(apply #'propertize "⚙ " `(face compilation-mode-line-run ,@props))))
|
|
≈ ALMOST EQUAL TO
≊ ALMOST EQUAL OR EQUAL TO
≅ APPROXIMATELY EQUAL TO
≃ ASYMPTOTICALLY EQUAL TO
≡ IDENTICAL TO
≙ ESTIMATES
≔ COLON EQUALS
|
|
|
|
How did I not remember this guy.
|
|
Would be neat if the constant face could be applied to this last
argument, just like it is highlighted in require and provide
statements. Alas, those are covered by font-lock rules in
lisp-mode.el…
|
|
The War For Mode-Line Space continues.
|
|
Footnote and MML stuff, mainly.
Change magit-blame-mode lighter: the new character is covered by Noto
Color Emoji, not the old one.
|
|
AFAICT, the last argument, ADD, will be nil by default,
dixit (elisp)Argument List. This changes the semantics since I used
to pass 'append explicitly, but I have no idea why I was doing that in
the first place, so 🤷
|
|
|
|
… an interesting game of whack-a-quote.
I thought ("..." . ?…) might work, but alas…
|
|
And start organizing them.
I don't really like this hanging '(, but I couldn't find a way to turn
it into something along the lines of (list …).
|
|
|
|
I use whitespace-mode to check out *whitespace*, not to got on a
crusade to enforce the Holy Eighty Rule.
|
|
|
|
This package seems to do exactly the kind of thing I would like to do.
|
|
|
|
Thereby making the function work with Emacs 25.
|
|
|
|
Get rid of electric-pair-mode binding; I can't remember the last time
I turned it off.
|
|
No idea how I let this one slide.
|
|
See linked bug report.
|
|
This is somewhat more informative for me.
|
|
|
|
|
|
It'd be really great to make Emacs grok Emoji faces.
|
|
|
|
|
|
I sometimes need to create a file named $(date +%F).md. This commit
introduces two ways to do this:
1. with recursive minibuffers:
C-x C-f M-! date +%F RET
2. with a new function to add shell command outputs to the kill ring:
C-x C-f C-c k ! date +%F RET C-y
|
|
- make magit-blame take less space in mode-line
- refine hunks, now that refinements look OK with eighters-theme
|
|
- use my/ prefix on every defun
- delight whitespace-mode
- remove leftover binding in C mode
- enable all "confusing" commands
|
|
I regularly used C-x C-f M-n C-SPC C-a C-w C-g to get the current
buffer's filename in the kill-ring (and the clipboard). This had two
issues:
1. It does not work with Ivy: C-a only moves back to the beginning of
the basename;
2. M-n runs the find-file-at-point machinery; if point is on a word
that looks like a domain (ends with ".com", ".net", …),
ffap *attempts to contact the domain*, which is inconvenient for a
bunch of reasons (locks up the editor, leaks data by sending it in
DNS requests, opens 9/tcp connections to random domains…)
The latter can be disabled by customizing ffap-machine-p-known; maybe
I'll go ahead and do that someday. In the meantime, defining a proper
function instead of relying on side-effects seems like a quick-win.
my/kill-ring-pipe-region allows me to quickly run pandoc on a Markdown
snippet and paste the resulting HTML in LWN's comment form. I'm sure
I will find other uses for it.
|
