Re: Implicit close of idle streams

On Wed, Aug 14, 2013 at 10:25 AM, James M Snell <jasnell@gmail.com> wrote:

> On Tue, Aug 13, 2013 at 3:46 PM, Martin Thomson
> <martin.thomson@gmail.com> wrote:
> > On 13 August 2013 23:38, James M Snell <jasnell@gmail.com> wrote:
> >> That part that is confusing the most I think is the "might have been
> >> initiated by that peer" part. The way I understood the state model,
> >> "idle" streams are not "initiated"... All streams start in the "idle"
> >> state. The word "initiate" is used in a number of places when talking
> >> about putting streams into the "open" state.
> >
> > Let's whittle this down a little: would s/initiated/opened/ fix this,
> > or is this something related more generally to the "might have" thing
> > (i.e., a client opening stream 5 implicitly closes streams 1 and 3,
> > but not streams 2 or 4).  Do you think that the latter needs more
> > clarification too?
>
> The part that is confusing is: What if stream 3 is open. The way the
> text currently reads, it's not clear if opening 5 causes open streams
> to close... vs. opening 5 causes all *previously unused* streams with
> lower stream ID's to be automatically closed, while leaving non-idle
> streams with ids < 5 alone. (I know how it's supposed to work, of
> course, but the way it's worded currently, it's not clear, and I'm
> afraid that it could be confusing and misleading, particularly to
> non-native english readers.
>
>
I'm non-native English readers and if you ask me it is confusing then I'll
answers yes.
My understanding is that that explains the monotonically increasing stream
identifier scheme
in formal way. I think it would be less confusing to convey that notion.


Best regards,


Tatsuhiro Tsujikawa