search

mnot / I-D

My Internet-Drafts
https://mnot.github.io/I-D/
Other
98 stars 38 forks source link

does available hints allow multiple default setting #349

Open Tanych opened 2 months ago

Tanych commented 2 months ago

For the available hints like AvailLanguage, I have a couple of questions to confirm:

  1. when the specs says "unrecognised Parameters on Structured Fields are ignored and therefore do not trigger this fallback", does it mean that e.g. en, zh;ddd, ja;dd should result in en, zh, ja, also en, zh;d;c and en, zh;d=?0 and en;d=1.0, zh should be treated the same as en, zh?

  2. whether multiple default setting allows or be ignored? it's not clear whether en, zh;d, ja;d should actually be rejected or be treated the same was as en, zh;d, ja or as en, zh;d.

recvfrom commented 1 month ago

Any thoughts on this? This came up in a Chromium code review and it'd be helpful to have a clarification to ensure that we've implemented parsing of the AvailLanguage header correctly. Thanks!

For reference: https://chromium-review.googlesource.com/c/chromium/src/+/5703436/comment/eac735bf_760f79bd/

mnot commented 1 month ago

The text in the spec needs some more nuance, but unrecognised Parameters on Avail-Language should indeed be ignored. If d has a non-Boolean value, we'll need to define (probably most sensible to ignore those too).

For multiple d, I'm inclined to say either first one wins, or it means the entire field should be ignored - thoughts?

recvfrom commented 3 weeks ago

Thanks Mark, and my apologies for the delay in getting back to you.

Ignoring unrecognized parameters and non-Boolean parameter values SGTM. I wonder if it'd also be helpful to have the spec say explicitly that the type of the d parameter is Boolean.

Regarding multiple 'd', I don't know if there is any precedent from how other specs handle this type of thing, but first one wins seems reasonable to me.

I've opened https://crbug.com/1456351 on our end to track making changes based on this feedback. Thanks again!