gnu-arch-users
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Gnu-arch-users] Re: Design proposal to kill pristine trees


From: Tom Lord
Subject: [Gnu-arch-users] Re: Design proposal to kill pristine trees
Date: Wed, 8 Oct 2003 18:53:37 -0700 (PDT)

    > From: Miles Bader <address@hidden>

    > > I don't think you want "a hook that replaces (or returns) a revlib
    > > path".

    > What I probably really want is `local' revlibs.  Having a hook that
    > returns a revlib path allows me to do this without getting into big
    > arguments over what the right algorithm is.

What algorithm?  What do you mean by "local revlibs"?

You do realize that tla does not currently add to a revlib except
under explicit direction, right?

In the future, perhaps that will change -- but if it does, you'll
certainly be able to override that behavior with a hook that adds the
revision to the revlib of your choice.


    > > I think you want a hook that notices demand for a locally cached copy,
    > > and hard-link algorithms that are sensative to devices.

    > Perhaps adding enough hacks in _other_ hooks might address the
    > hard-linkability issue (though I'm not convinced of that yet),


_What_ issue?  If some code wants to build a link tree, and it has a
choice of revlibs, it should look for one that it can actually make a
link tree from.


    > of the above address the `manageability' issue -- global resources get
    > to be a pain beyond a certain point.  It's like the comment about
    > pristines being nice under {arch} because they go away when you remove
    > the tree; I want something a little different, but for similar reasons.

?


    > I'd like to keep things localized in some way; 

?

-t




reply via email to

[Prev in Thread] Current Thread [Next in Thread]