To remove the ~(require 'cl)~ dependency, I wrote my own reduce for
boolean values, recursive, which should do the trick. Mostly just to
remove the annoying "haha no cl library anymore".
Set to 4 workers at the start to make compilation at least as fast as
possible: my laptop has 4 cores so it'll just slow it down until early
compilation is done, while my desktop can still keep going as it has 8
threads. In config.org this is setup correctly so the rest of the
config (which is the much larger part) uses a system dependent number
of cores.
When emacs is launched standalone, it makes sense to optimise towards
smaller load times through lazy loading. However, when Emacs is in
server or daemon there's no point holding back. Just load everything
most of the necessary stuff such that on first emacsclient launch we
have everything available for use.
This is a smarter Emacs literate config. The way it works is simple:
- If no config output files exist, compile them. Otherwise just load
the config
- Add a hook when exiting emacs such that the config is compiled and
ready for next call.
This makes it so that load time is fast and compile occurs at a more
opportune time than at init.
This one is a lot more closer to Emacs core, it doesn't use general
or the SPC leader as I'm used to but instead the default meta-key
chord bindings to work better. I don't use a lot of features in Emacs
that Doom used to provide in bindings, so I'm stripping them away and
slowly adding stuff I like.
Developed quite a bit of the config.
I've added a literate org config, loaded through the init.el file.
Furthermore, I've added some basic general maps and have gained a much
better understanding of Emacs than before. This refactor is one of the
main reasons I decided to so heavily refactor my doom config.
+better binding system (general) with a leader
+evil-surround
+which-key
+org
+counsel
and tango-dark cos if I see default emacs again I'm going to throw my
computer outside.
