[pacman-dev] [PATCH] Add new SharedPkgCache option.

Roman Kyrylych roman.kyrylych at gmail.com
Mon Feb 18 03:06:35 EST 2008


2008/2/18, Dan McGee <dpmcgee at gmail.com>:
> On Feb 17, 2008 10:14 AM, Roman Kyrylych <roman.kyrylych at gmail.com> wrote:
> > 2008/2/17, Xavier <shiningxc at gmail.com>:
> > > On Sun, Feb 17, 2008 at 05:37:13PM +0200, Roman Kyrylych wrote:
> > > > 2008/2/16, Chantry Xavier <shiningxc at gmail.com>:
> > > > > As it was already mentioned several times, the new -Sc behavior in 3.1 is
> > > > > great, but only when the package cache is not shared.
> > > > >
> > > > > When this option is enabled, -Sc will clean packages that are no longer
> > > > > available in any sync db, rather than packages that are no longer in the
> > > > > local db. The resulting behavior should be better for shared cache.
> > > > >
> > > >
> > > > Great name! :-)
> > > >
> > > > BTW, I've just realized that there might be a problem with "no longer
> > > > in any sync db" that, I think was not in 3.0's -Sc:
> > > > if SystemA has core,extra,community enabled and SystemB only
> > > > core&extra then doing -Sc on SystemB will cleanup community packages
> > > > from the shared cache.
> > > > I don't know if it's worth to fix this by using a more complicated
> > > > (filename comparison, AFAIR) aproach like in 3.0. :-/
> > > >
> > >
> > > Indeed, that's a problem, but don't know if it's a big one. I would think
> > > it's rather common to use the same set of sync db on a set of machines
> > > sharing pkg cache. And that approach is indeed much simpler.
> > >
> > > I mentioned this in the man page, that this option is rather meant for setup
> > > where same sync dbs are used.
> >
> > Ah, OK, didn't read that. :-)
> > Then we can forget about this corner case,
> > it can be worked around by cache cleanup scripts based on filename parsing
> > (available on BBS and ML).
>
> And now we are back to the whole reason why I think this is a
> pointless problem to solve. :P

Not pointless at all, Xavier's solution will work on majority of
configurations with a shared cache (that have the same set of repos
enabled).

>
> Once you venture into the shared cache realm, I just don't see a need
> for pacman to hand-hold you.

It's useful for users that don't use a shared cache, but want to keep
non-installed but up-to-date packages in cache (e.g. me :-P) - there
are some valid reasons for doing this.

>
> So we have Xav and Roman in support of this, and me against it. I
> don't want to say "majority rules" here, but I would like to get a few
> more voices on this one. Aaron?



-- 
Roman Kyrylych (Роман Кирилич)


More information about the pacman-dev mailing list