update README for pacman CacheServer option

[HarHarLinks]

This, or the mirrorlist part of it, is what I have been running for half a year now, works splendidly.

Resolves #104

[haydonryan]

@anatol please merge this - This should be the "default" way. it's super important as it fixes one of the issues I had with pacoloco - running on a laptop that's not always connected to my home network (and therefore pacoloco was not accessible.)

Keep up the great work!

[daurnimator]

Cache servers will be tried before any non-cache servers, will not be removed from the server pool for 404 download errors, and will not be used for database files.

I'm not sure we want a CacheServer directive: pacoloco handles database files.

[haydonryan]

Cache servers will be tried before any non-cache servers, will not be removed from the server pool for 404 download errors, and will not be used for database files.

I'm not sure we want a CacheServer directive: pacoloco handles database files.

I'm not sure I follow, can you elaborate please?

[HarHarLinks]

See https://gitlab.archlinux.org/pacman/pacman/-/merge_requests/98#note_133948 though I'm not sure what the security concerns mentioned are to build an opinion based on that aspect.

Personally I don't really care about fetching 3 database files but the 400 packages that are worth caching because they're either big (bandwidth) or tiny (latency). My use case with this is for my laptop to use the cache at home or another location where the cache is reachable, but when not reachable don't take a lot of time to error out etc before skipping to the next mirror.

From pacman POV, this seems to be the intended/recommended way to use a caching proxy such as pacoloco.

[haydonryan]

Super helpful thanks!
I think it depends what you value - if you want the option of fetching packages off your own network eg laptop usecase, then cache server works.

If you want your pacoloco to be always up and only ever pull from that, then use the replace servers approach.