Think this is done in Kirby 3
1 Like
No, not in Kirby 2…
Unfortunately not (and some more chars)
You could test with your own cache driver if removing and recreating the cache folder is faster than the current method.
https://getkirby.com/docs/developer-guide/advanced/caching/#cache-drivers
Do you have a hint what the “nuke folder” command is on linux? AFAIK rm -rf recursively deletes the contents first too before removing the directory. I’m afraid it’s not faster then find -delete…
I just tested what happens when the cache directory is not present, as in theory this can happen when I’m recreating it. Then I get a HTTP ERROR 500 from the server, which is not a good solution imho.
I decided to fix it quick and dirty by running a nightly cronjob which flushes cache (via rsync - it looks like its a tiny bit faster) as @texnixe suggested. 
1 Like
bit late to the party.
for k2 and the big sites i created i never used kirbys build in cache but split the site into cachable parts and just aggregated the results. i used the kirby cache folder and a 2-monthly cronjob.
i mainly used my own implementation or https://github.com/kirby-deprecated-plugins/kirby-time-cache.
for k3 i am using my https://github.com/bnomei/kirby3-lapse.
the tricky part is to make the parts small enough to track modification correctly but also efficiently (check & loading must be faster than creating). what i also did was making the cache alway-write-only when in debug mode – making it easy to force updates.
my usecase was creating a headings list for 3000 pages with thumbs and similar stuff. but i must say that in most other cased the built in cache and a regular crobjob should be fine.