Re: Deadlocking in the transport

[Jana Iyengar <[e-mail]>]

>
> Ahh, excellent, then I think at least we two are in agreement.  It
> seems like there is an emerging acceptance of the problem and proposed
> approach, I guess we might start considering how to address it.
>

I think so. I've heard some other approaches aired that folks may want to
discuss however.


> I think that this needs to be in the main spec.  Failing to document
> this sort of pitfall could be fatal.  Does anyone disagree?
>

I agree that we should document this, especially since it's expected to be
a common design pattern. Since we don't express dependencies among streams,
we have to ensure that an app that wishes to do so doesn't do the wrong
thing. I think we definitely should draw attention to this point, but I'm
not sure that we need to be very prescriptive here for what the app should
do, but I think we should use strong language for what a transport
implementation should do.

We should also mention that an app that chooses to not read out of a stream
buffer may cause negative interactions with any autotuning mechanisms
implemented to tune window update magnitude and frequency.