Re: Use cases for local-end set session id?

On 4 Feb 2015, at 19:32, James Graham <james@hoppipolla.co.uk> wrote:
> What's the use case for local ends requesting a session id? As far as I
> can tell from the spec at the moment, the local end and all
> intermediaries have to deal with the possibility that the remote sets
> its own session id that's different from the requested one. Therefore
> this feels like a feature that adds a lot of implementation complexity
> and high possibility for bugs (e.g. because code assumes that it will
> always get the session id it requested, but doesn't), but has little
> value since it can't actually be relied on. Therefore my a-priori stance
> is that it should be removed.

If there are no proponents for this added complexity, I suggest we move to remove it.

Received on Monday, 9 March 2015 08:39:26 UTC