CONTINUATION frame and continuation flag

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

CONTINUATION frame and continuation flag

Alexey Melnikov
Currently in draft -04 we have:

In "HEADERS" frame:

   END_HEADERS (0x4) flag

In "PUSH_PROMISE" frame:

   END_PUSH_PROMISE (0x1) flag


I am updating the document to introduce a new CONTINUATION frame as per
issue 183. I have a couple of questions:

1) do people want to keep using END_HEADERS/END_PUSH_PROMISE flags or
should I replace them with "more headers to follow" flag?
(Another option is to remove the flags altogether and just derive this
information from presence of CONTINUATION frames. Speak up if you prefer
this option.)

2) If the answer to 1) is "more headers" flag, can I reorder existing
flags on HEADERS/PUSH_PROMISE so that the new flag
has the same value for both frames (and the new CONTINUATION frame)?

Comments?


Reply | Threaded
Open this post in threaded view
|

Re: CONTINUATION frame and continuation flag

David Morris-4

Keep the binary value for the flags the same ... either all last/end=1 or
all more/continued=1. I prefer the former value, but not very strongly.

On Tue, 6 Aug 2013, Alexey Melnikov wrote:

> Currently in draft -04 we have:
>
> In "HEADERS" frame:
>
>   END_HEADERS (0x4) flag
>
> In "PUSH_PROMISE" frame:
>
>   END_PUSH_PROMISE (0x1) flag
>
>
> I am updating the document to introduce a new CONTINUATION frame as per issue
> 183. I have a couple of questions:
>
> 1) do people want to keep using END_HEADERS/END_PUSH_PROMISE flags or should I
> replace them with "more headers to follow" flag?
> (Another option is to remove the flags altogether and just derive this
> information from presence of CONTINUATION frames. Speak up if you prefer this
> option.)
>
> 2) If the answer to 1) is "more headers" flag, can I reorder existing flags on
> HEADERS/PUSH_PROMISE so that the new flag
> has the same value for both frames (and the new CONTINUATION frame)?
>
> Comments?
>
>

Reply | Threaded
Open this post in threaded view
|

Re: CONTINUATION frame and continuation flag

Martin Thomson-3
On 6 August 2013 15:33, David Morris <[hidden email]> wrote:
>
> Keep the binary value for the flags the same ... either all last/end=1 or
> all more/continued=1. I prefer the former value, but not very strongly.

Yes, we agreed to this before and I don't think it's worth backsliding
on this one.  END_CONTINUATION >> NOT_DONE_YET.

Reply | Threaded
Open this post in threaded view
|

Re: CONTINUATION frame and continuation flag

James M Snell
+1 ...

On Tue, Aug 6, 2013 at 7:31 AM, Martin Thomson <[hidden email]> wrote:
> On 6 August 2013 15:33, David Morris <[hidden email]> wrote:
>>
>> Keep the binary value for the flags the same ... either all last/end=1 or
>> all more/continued=1. I prefer the former value, but not very strongly.
>
> Yes, we agreed to this before and I don't think it's worth backsliding
> on this one.  END_CONTINUATION >> NOT_DONE_YET.
>