Re: END_PUSH_PROMISE

+1 ... meant to ask about that possibility (1 END bit). If this were ages
ago in time, I'd vote for 0x1 as the common bit, but that seems likely
to be more disruptive than useful.

On Wed, 21 Aug 2013, Martin Thomson wrote:

> Alexey was a little coy about raising this issue.  -05 has
> END_PUSH_PROMISE as 0x1, whereas END_HEADERS and END_CONTINUATION are
> both 0x4.
> 
> I've changed END_PUSH_PROMISE to be 0x4 on the grounds that
> consistency is easy to attain at this stage and there is no other
> reason for this to be 0x1 other than that being the least significant
> bit.
> 
> Scream if this causes you some sort of pain.  I'll revert the change
> and we can discuss.
> 
> (We still haven't resolved the issue whereby a continued PUSH_PROMISE
> can end the stream it's associated with, I've just added an editors
> note to the draft on that point.)
>

Received on Wednesday, 21 August 2013 20:47:08 UTC