ietf-wg-wimse / draft-ietf-wimse-s2s-protocol

WIMSE Service to Service I-D
Other
3 stars 3 forks source link

Why ABNF for WIT? #32

Open yaronf opened 2 months ago

yaronf commented 2 months ago

In Sec. The WIT HTTP Header, we define the ABNF even though we say the WIT is a JWT and this completely defines the syntax. IMO the ABNF could actually confuse implementers.

bc-pi commented 2 months ago

The last time I went through the process of requesting registration of an HTTP Field Name to carry a JWT, there was some grumbling from one of the designated experts about it not being defined as an RFC 8941 Structured Field Value. When it was explained that there wasn't really an appropriate SF type to carry a JWT (and especially that sf-binary didn't fit at all) the grumbling shifted to general complaints that the syntax wasn't well defined (despite, as you point out, the syntax of a JWT being completely defined by JWT). The grumbling was partially mollified by the addition of some less than useful ABNF (found here). It was more involved than that, of course, but that's the basic story line. So I thought I'd try and get ahead of that kind of DE push-back a bit while adding some ABNF that's a little more descriptive/useful. Which is what is here. I think the WPT header field should also get some ABNF but I couldn't fit it in as well and ran out of time for the pre -00 PR deadline.