What to do about HPSS protocol backwards compatibility for faster push?

Robert Collins robertc at robertcollins.net
Wed Oct 17 22:05:01 BST 2007


On Wed, 2007-10-17 at 22:54 +0200, Vincent Ladeuil wrote:
> >>>>> "Andrew" == Andrew Bennetts <andrew at canonical.com> writes:
> 
>     Andrew>   3) send a “do you support request X?” request first
>     Andrew>   (possibly integrate this into the existing “hello”
>     Andrew>   request).
> 
> +1
> 
> I thought smart methods embedded a version too, or was it
> abandoned ?
> 
> Initial handshake should provide the client version to the server
> and returns the server version to the client.
> 
> From there it should remains no ambiguities about who is able to
> do what.
> 
> Or did I miss something obvious ?

It would be nice to remove the existing hello; its a stateless protocol
so its technically possible to do:
HELLO
ANSWER Version X
--version change--
STREAMDATA
ERROR-NO-SUPPORTED

But more importantly, the hello is a round trip that adds to the minimum
time for operations such as no-op pushes, and getting rid of it is a
goal.


-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/20071018/3e73ddb6/attachment.pgp 


More information about the bazaar mailing list