Re: Comments on draft-ietf-secsh-filexfer-05.txt

[Damien Miller <djm%mindrot.org@localhost>]

> While parsing a packet was a straightforward, low-layer activity, it is 
> now necessary to manage the information about sent packets.

Surely any application will need to hang onto this information anyway -
how else would it match replies (which may appear out of order*) with
requests?

-d

* I'd prefer if this behaviour was banned by the spec, it adds
complexity to client implementations and makes things like resumable
transfers more difficult to reliably implement.