- From: Martin Thomson <mt@lowentropy.net>
- Date: Mon, 08 Dec 2025 12:49:42 +1100
- To: "Julian F. Reschke" <julian.reschke@gmx.de>, art@ietf.org
- Cc: draft-ietf-httpbis-incremental.all@ietf.org, ietf-http-wg@w3.org, last-call@ietf.org
Hi Julian, https://github.com/httpwg/http-extensions/pull/3349 contains changes in response to your helpful review. On Sun, Dec 7, 2025, at 04:24, Julian Reschke via Datatracker wrote: > In 3., two documents about "extended CONNECT" are mentioned. Both are about > WebSockets. Would it make sense sayint it here, potentially sparing readers the > indirection? Unfortunately, extended CONNECT was defined -- in both cases -- in documents that also defined WebSockets use of extended CONNECT. But this is a generic capability, not specific to WebSockets, so it might be a bit misleading to mention WebSockets. > 4.3: it's (to me) not complely clear what is recommended. Set limits on bytes > or time? The subsequent sentence indicates that this would be a problem as well. Good point. We need to be clearer. If you buffer, then you can set a limit in bytes (when I have X bytes, continue) or time (after X milliseconds, continue). The former condition is not guaranteed to ever trigger, whereas the latter is, so we should insist that a limit in time be mandatory, a limit in bytes is optional. > IANA: maybe link to the HTTP Proxy Error Types registry - that would make it > easier to check the requirements. Done for both.
Received on Monday, 8 December 2025 01:50:07 UTC