Branch uses Transport and Storage objects

[John A Meinel]

John A Meinel wrote:
> I've been working on updating the internals of the Branch class, so that
> it can support multiple protocols using a pfs-like abstraction layer.
>
> It is available here:
> http://bzr.arbash-meinel.com/bzr-split-storage/
>

...

> Please give it a look over. I'm hoping to have all the tests passing soon.

Now, as of revision 934, I have all tests passing, and I'm working
through some of the issues with remote branches (since these are not
covered by the test suite).

The only portion I'm not sure of, is that I have a Branch object
automatically create a local cache if transport.is_remote() evaluates to
True.
Which means that the Branch needs to cleanup it's cache when it goes
away. And I know past discussions decided that having a Branch object do
cleanup in it's deconstructor was bad form, because you cannot guarantee
it will be cleaned up properly.

So would it be better to have an explicit Branch.setup_cache(), which
means that the calling function will be aware and clean up the cache
root in a try/finally.

I just would like to make it easy to do caching, so that commands that
can connect to remote branches don't have to think about it too much.

John
=:->
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 253 bytes
Desc: OpenPGP digital signature
Url : https://lists.ubuntu.com/archives/bazaar/attachments/20050711/e87f5ad3/attachment.pgp