Re: END_SEGMENT and headers

I misread that!

I withdraw that this is a good improvement, since there are protocols that
may want to have metadata in the middle of a message.

-=R


On Tue, Apr 15, 2014 at 10:00 PM, Roberto Peon <grmocg@gmail.com> wrote:

> Seems like a reasonable reduction/improvement to me.
> -=R
>
>
> On Tue, Apr 15, 2014 at 9:54 PM, David Krauss <potswa@gmail.com> wrote:
>
>> The END_SEGMENT flag is defined for HEADERS frames, but it seems to be
>> meaningless. A literal reading of the draft seems to say that such a frame
>> cannot be coalesced with subsequent CONTINUATION frames, but those frames
>> do not also have the same (useless) segmentation capability.
>>
>> Why not let END_SEGMENT and END_HEADERS be the same bit? Then headers be
>> treated as the first segment of any stream, by some simplistic
>> intermediaries. It also reclaims a flag bit, as they are running into short
>> supply.
>>
>> Also, is END_SEGMENT required when END_STREAM is set? It seems it should
>> be, for the sake of protocol regularity.
>>
>>
>>
>