- From: Willy Tarreau <w@1wt.eu>
- Date: Fri, 21 Jan 2022 10:00:06 +0100
- To: Martin Thomson <mt@lowentropy.net>
- Cc: ietf-http-wg@w3.org
On Fri, Jan 21, 2022 at 05:42:09PM +1100, Martin Thomson wrote: > Please, take a quick look and tell me if you see anything that worries you. Looks very good overall. I noticed one nit in 6.5.3: "some allowance need to be made ..." It seems to me that "allowance" is singular hence we ought to have "needs to" here, but I'm not a native speaker and could be wrong. Another point, I remember that during my early implementation I was in trouble trying to figure how to perform a late stream close after a transfer (e.g. headers only, or end of data detected). I was advised that the normal way to do this was to send a zero-length DATA frame with the ES flag set in this case. It's not obvious at all that it's the normal way to do it, especially when dealing with methods or statuses that do not take data. The current text about the DATA frame speaks a lot about padding tricks but never mentions anything about zero length being valid. There's a useful note at the end saying "A frame can be increased in size by one octet by including a Pad Length field with a value of zero". I think the spec would benefit from one extra such note saying: "A late end of stream may be signaled by sending a DATA frame with a length of zero octets and the END_STREAM flag set". Otherwise everything looks pretty nice to me :-) Willy
Received on Friday, 21 January 2022 09:00:28 UTC