Re: draft-ietf-httpbis-http2-latest, 6.2 HEADERS, 6.10 CONTINUATION, END_SEGMENT (0x2)

> On 3 July 2014 22:20, Kari Hurtta <hurtta-ietf@elmme-mailer.org> wrote:
> > So now it is fixed.
> 
> 
> Thanks.  I wasn't sure if that would help.

Yes, that fixed that first "item" on my mail:

| What I have missed ?
| 
| How END_SEGMENT is supposed to work when segment is supposed to end
| after headers and headers do not fit to one HEADERS frame ?


( If any new flags are added to HEADERS, then these flags
  also must be specified are they for that HEADER -frame
  or are they for combined HEADER + CONTINUATION frames. ) 


You quoted that another part of mai mail:

| What "A HEADERS frame that is followed by CONTINUATION frames carries
| the END_STREAM flag that signals the end of a stream." that is exactly
| supposed that say:

That was partially misread from my side.



Yet about that first "item" on my mail:


That

| END_SEGMENT (0x2):
|
|     Bit 2 being set indicates that this frame is the last for the current segment. 

should probably be

>    Bit 2 being set indicates that this frame (and possible CONTINUATION frames)
>    are the last for the current segment. 

That part

|    END_SEGMENT, like END_STREAM, applies to a complete sequence of HEADERS and 
|    CONTINUATION frames. 

of course implies that but in that situation "this frame is the last for the 
current segment" is not exactly true.



Alternative that han be said same way than on "END_STREAM (0x1)" part:

> END_SEGMENT (0x2):
>
>    Bit 2 being set indicates that this header block is the last for the current 
>    segment.


That way there is not "this frame" but "this header block", which is easier
to understand as HEADER frame + CONTINUATION frames.


/ Kari Hurtta

Received on Monday, 7 July 2014 09:40:11 UTC