archives & nested trees

[Martin Pool]

On 14 Nov 2005, John A Meinel <john at arbash-meinel.com> wrote:

> What I feel we are really missing is the concept of archive names, which
> can then point to whatever you have registered. Since right now, the
> full archive path is its name, and thus we don't have a way to point it
> to somewhere else.
> There has been some discussion about using aliases, both for shortening
> the full path, but also to say that "foo wants to go to path/X, but I
> know that path/Y can be substituted for path/X".
> The substitution helps if someone publishes an archive, and says go to
> "sftp://somewhere/path/to/here" and I know that I really want to get it
> at "http://somewhere/here".

I'd suggest doing it the other way around: make the long-term publicly
readable value be the canonical name for the branch.
 
> If we want to go back to named archives, with named mirrors, we can. But
> I'm certainly not settled that that is the best way to go. (For example,
> it makes having standalone branches difficult).
> We could make the naming optional. Where everything can be referenced by
> full path, but if you happen to have an archive named, it can also be
> referenced by that.

Yes, I think this may be the best way to do it.  To my mind the problem
with arch archive naming is not so much the names, but rather that they
propagate so strongly into revisions which are created within them, so
causing trouble with reuse of archive names, laptop mode, etc.

If the included-by-reference branches use aliases then it's reasonable
to expect people to register them before building.

-- 
Martin
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
Url : https://lists.ubuntu.com/archives/bazaar/attachments/20051117/ec7b0057/attachment.pgp