HttpTransport with async pipelining

[Aaron Bentley]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Martin Pool wrote:

> Something in the latest merge code (run by bzr branch) seems to have
> trouble with accessing branches that have a non-default root id.  I get 
> 
> bzr: Changeset entry for TREE_ROOT-20050718050618-bed150f7b5473c3a (./.)
> is invalid.
> file ./. has "TREE_ROOT-20050718050618-bed150f7b5473c3a" id for both
> self id and parent id.
> 
>   at /home/mbp/lib/python/bzrlib/changeset.py:557 in __init__()
>   see ~/.bzr.log for debug information
> 
> Maybe that check should just be relaxed for the root?

I don't think so, because the root's parent is not supposed to be
itself.  No file should have itself for a parent.  In the merge code,
the ID of the root's parent was changeset.NULL_ID, i.e. "!NULL".
Consistency with bzr would suggest using None as the parent ID, but self
doesn't make sense to me.

I am trying to reproduce this problem, but not having much success.  Any
suggestions?

Aaron
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFC9snQ0F+nu1YWqI0RAusoAJ920U0922MCz6zYNLTCYQYMT66+2QCfd155
mypThhdwOxbwJTwjAAPdBEQ=
=Yf9C
-----END PGP SIGNATURE-----