Re: Who is using experimental codepoints in HTTP/2?

--------
Martin Thomson writes:
> https://github.com/httpwg/http2-spec/pull/853 proposes that we do away with t
> he "Experimental Use" ranges for frame types and settings.  The result would 
> be that new registrations could use these values.

I doubt this will cause issues for H2, even with a 8-bit space.

But it does give me reason to pause if you want this to be the general
policy going forward ?

Either we have to size fields which index such "potentially unlimited"
registries generously to avoid any "X bits is enough for everybody"
embarrasment, which probably means >=32 bits or we have to serialize
them as variable length.

SNMP did the latter and that seems like a good call: The
`enterprise-numbers` IANA registry is now up to 57468, despite the
people advocating for 16 bit fields back then assuring us there
would be /no way/ for that to ever happen.

-- 
Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk@FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.

Received on Monday, 24 May 2021 15:06:10 UTC