nested tree remaining questions

[Robert Collins]

On Tue, 2006-02-07 at 21:43 -0500, Aaron Bentley wrote:
> |  * What child location data do we need? I think all we need is
> | rootid:url pairs. We can of course try all the urls, but the root id is
> | a nice hint. multiple entries per url and per rootid are allowed, to
> | give fallback facilities and the like. Seeking a +1 on this.
> 
> I'm not really clear on what this is for.  Wouldn't it make more sense
> to associate the data with the nested trees themselves?  And if the
> nested tree has a branch, should it be updated?

what it is for:
So, imagine I have two branches, parent and child. Both are standalone
branches - there is no repository involved. I publish parent at
ROOT/parent/trunk and child at ROOT/child/trunk.

A bzr get/branch/clone of ROOT/parent/trunk needs to retrieve
ROOT/child/trunk automatically. 

We cant put the data in the child tree, because its data *about* child
that we need.

We could put it in the inventory entries content for the child, but that
seems somewhat messy as its then harder to find places to try when there
is no hint for a given root it.


The algorithm for 'update' in the parent tree that has a child nested
tree is to do an 'update -r revid' in the child tree, where revid is the
new revid in the parent trees inventory entry for the child tree. The
source branch that revisions are pulled against is the childs parent, or
if that is missing the needed revisions, we consulting the
child-locations table in parent.

Rob


-- 
GPG key available at: <http://www.robertcollins.net/keys.txt>.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
Url : https://lists.ubuntu.com/archives/bazaar/attachments/20060208/f9448413/attachment.pgp