Thanks Felix. Draft 11 added the Brotli RFC7932 as a normative reference which describes the stream decoding algorithm and how to handle invalid streams (in addition to the security considerations).
We can add some language to this draft as well if you think it will help, but it's an extension of the existing brotli format and relies on the same underlying processing.
>
> I wanted to weigh in with my support for publishing this document. It is
> useful and informative.
>
> Over the years since it was first circulated, I have referred to this
> document a number of times in my work with dictionary-based compression.
>
> I have since given it a more thorough read, and I haven't found any
> issues that block its publication.
>
> In my recent reading, I have only one note:
>
> Section 1.4, "Compliance", describes how this document requires
> implementations to produce and consume messages conforming to this spec.
> It does not in general describe the behavior an implementation should
> have when it handles a message that does not conform to the format.
>
> This topic is touched on in RFC7932 Section 12 ("Security
> Considerations"), which is included by reference in this document's
> Security Considerations. Maybe that's sufficient, along with the various
> behaviors specifically described as elements of the format are defined.
>
> Best,
> Felix
-- last-call mailing list -- last-call@xxxxxxxx To unsubscribe send an email to last-call-leave@xxxxxxxx
