Re: Migrating a spec to structured fields

Roberto,

Thank you for providing this transition document, I think that collecting these will be helpful. I went through a similar exercise formally defining the new header format for Message Signatures in this pending PR:

https://github.com/httpwg/http-extensions/pull/1411 <https://github.com/httpwg/http-extensions/pull/1411>

We don’t have a full ABNF in there yet, but I’m seeing that we should add that in a future PR now and this is a good example of how to incorporate the fields and formats. We’ll have similar restrictions, like the expiration being non-negative.

And from an engineering perspective, I also pulled together both a client and server implementation of the signatures draft based on that PR, and using a Structure Fields library from Julian I was able to get it together in relatively short order as a result. Having a deterministic serialization for structured data is REALLY helpful for a security standard, so it’s been good to move in that direction.

 — Justin

> On Mar 9, 2021, at 11:53 AM, Roberto Polli <robipolli@gmail.com> wrote:
> 
> Hi folks,
> 
> I wanted to migrate the RateLimit headers with roughly the following abnf to S-F
> 
> ```
> RateLimit-Limit = expiring-limit [, 1#quota-policy ]
> expiring-limit = request-quota
> quota-policy = request-quota; "w" "=" time-window *( OWS ";" OWS quota-comment)
> quota-comment = token "=" (token / quoted-string)
> request-quota = quota-units
> quota-units = 1*DIGIT
> time-window = delay-seconds (from Section 10.2.4 of [SEMANTICS])
> ```
> 
> I stubbed a commentable doc here
> https://docs.google.com/document/d/1e9Vxcmb6hrle8LNKMtQc_eB-p7jdEPrBE1lgzrAqYKo/edit
> for all the willing people that are happy to provide hints.
> This could become a useful migration guide/reference to foster S-F adoption.
> 
> Thanks for your help,
> R.
>