Re: [Gnu-arch-users] Re: Mirrors and remote repositories...

[Tom Lord]

    > From: Greek0 <address@hidden>

    > On Sat, 06 Mar 2004 14:41:35 GMT
    > "David A. Wheeler" <address@hidden> wrote:

    > > In my opinion, a "commit" against a mirror should be automatically
    > > redirected to the mirrored repository.  Would that be hard to do?

    > I still don't really see the need for it. Where is the catch with
    > branching into your own archive and using star-merge?

As I recall, David is in a situation where he really just needs a
single "logical" archive -- wich needs to be remote, but there are
huge performance advantages for caching that archive locally for the
purpose of reads.

I agree with you that "Awe, suck it up and branch" is a perfectly fine
generic answer to his need however ---

The fact is that the general idea of a "write past mirror" fits in
hella cleanly with the code and the architecture.  Arch is entering a
stage of life where much of the development is about _detailing_ --
adding little curly-q's here and there that hit this or that niche
requirement.

David: I don't think that commts against mirrors should be
_automatically_ redirected.  Rather, I'd like to see a mechanism so
that you can separately register the read and write halves of a single
archive name.  The next question then is what to do about
auto-updating the mirror during reads.

-t