Arch/Baz import script

[Erik Bågfors]

On 4/30/05, Aaron Bentley <aaron.bentley at utoronto.ca> wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Robert Collins wrote:
> > On Thu, 2005-04-28 at 23:15 +1000, Aaron Bentley wrote:
> >>There's more to do, (more metadata, etc), but I've imported
> >>a small branch using it.
> >
> > Good starting point ;). What I'd want to be able to do is parallel
> > imports of the common history and have them identical in terms of bzr
> > inventory/log ids etc.
> 
> Inventory is already taken care of.  The Arch inventory IDs are copied
> into file IDs verbatim.  The revision IDs need a little more care,
> because a revision ID should reflect:
> 1. The fact that they're Arch revisions
> 2. The Arch revision ID
> 3. The algorithm used to import the revision
> 
> 3. is important, because it's conceivable that more than one approach
> could be taken to importing Arch revisions.  If it were omitted, it
> would be impossible to distinguish between two bzr branches derived from
> the same Arch branch, but with different contents/metadata due to a
> different import algorithm.

This kind of thinking is great.  How can this apply to cvs and/or svn?
 The whole thing with IDs are alot different.

Basically what I want to know if if two different bzr branches of
something that comes from cvs/svn can be distinguished between.

/Erik