Content-MD5

[mnot]

The deprecation of Content-MD5 in 7231 was a bit too subtle; it's only mentioned in a change note, and the registry wasn't updated.

[reschke]

How exactly should we update the registry?

[mnot]

change its status there to obsoleted, as it has for other selected headers.

[reschke]

History: https://trac.ietf.org/trac/httpbis/ticket/178

We removed the header field, which is different from deprecation (which we did for a new status codes). The reason might be that we hoped that a replacement for Content-MD5 would surface, and that spec would then obsolete C-MD5.

Can't we just instruct IANA to update the registry? It would be weird to talk about Content-MD5 in the new spec, when the previous one didn't have it (except in the "changes" section).

[mnot]

We can try, or we can put it in IANA considerations only.

[mnot]

Proposal:

Add to the table in 4.1:

| Content-MD5 | http | obsoleted | RFC7231, Appendix B |

[reschke]

My proposal is to leave the spec alone and just tell IANA.

[royfielding]

+1 to just telling IANA.

[mnot]

OK, will bring it up on the mailing list and see how that goes.

[mnot]

Brought up on the ietf-message-headers list: https://mailarchive.ietf.org/arch/msg/ietf-message-headers/ngv6EIWfWjJCIR3ilxDn8OEH38w

[aamelnikov]

IANA/IESG prefers for the action to be recorded somewhere. So I think mentioning this in a future RFC would be the best.
If people think that this is urgent, I can ask IESG approve obsoletion.

[mcmanus]

I think its fine to do this in the IANA considerations of http-core itself. perhaps with 1 sentence explaining the situation.

…

On Wed, Nov 14, 2018 at 9:30 AM aamelnikov ***@***.***> wrote: IANA/IESG prefers for the action to be recorded somewhere. So I think mentioning this in a future RFC would be the best. If people think that this is urgent, I can ask IESG approve obsoletion. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#93 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAP5s09ZXfaAV9fBudPwp5wz6GUryi5rks5uvFNIgaJpZM4UsmsW> .

[mnot]

As per discussion in #188, should this be deprecated, not obsoleted?

[royfielding]

As Julian said, if we still define the header field it is deprecated; if we don't define the header field (meaning its only definition is in obsolete documents) then it is obsoleted.