Edge local fingerprint looks like SDP ones

[ibc]

In Edge latest version, the fingerprints values of iceGatherer.getLocalParameters() have a value such as "75:1B:81:93:B7:ED:27:7E:42:BE:D6:C4:8E:F7:04:3A:49:CE:3F:EE" rather than "751b8193b7ed277e42bed6c48ef7043a49ce3fee".

Should I understand that this is a bug in Edge? or should implementations be ready to deal with both syntaxes (by detecting usage of ":")?

[fippo]

It has been like that forever. The spec says

The value of the certificate fingerprint in lowercase hex string as expressed utilizing the syntax of 'fingerprint' in [RFC4572] Section 5.

which, if you go to https://tools.ietf.org/html/rfc4572#section-5 is

The hash value is represented as a sequence of uppercase hexadecimal bytes, separated by colons

It is somewhat odd to require lowercase while 4572 says uppercase.

[ibc]

Thanks. And sorry, I meant RTCDtlsTransport. So it seems the ORTC spec is wrong regarding this. What about mandating uppercase separated by colons and that's all?

[robin-raymond]

The specs themselves are inconsistent. We went back and forth on this one... I personally favour JS style lowercase byte arrays without : but to make the shims easier we followed the spec. The trouble is one spec mandates lower case : and another uppercase : and the examples/samples are inconsistent. Sadly, I think it's best clients/implementations handle any of these cases, with or without :.

[ibc]

The trouble is one spec mandates lower case : and another uppercase :

Well, what I see is that one spec mandates lowercase without : by wrongly referring to another spec that mandates the opposite.

[robin-raymond]

@ibc no, the two specs I'm referring to are not the ORTC spec. There are two "related" specs with differing versions.

[ibc]

Fine but, anyhow, this is wrong:

The value of the certificate fingerprint in lowercase hex string as expressed utilizing the syntax of 'fingerprint' in [RFC4572] Section 5

This is just about removing such a reference to RFC 4572.

[robin-raymond]

@ibc I agree and would say if we are mandating it look like this RFC with the : then it should be uppercase like this specification despite other conflicting specifications that exist since this is likely the most authoritative of the ones we want to model.

[ibc]

Agreed. Also, AFAIK, there are some case-sensitive implementations, so forcing it to be always uppercase seems to make things easier.

[lgrahl]

I'd also upvote uppercase.

[aboba]

I vote for uppercase as in RFC 4572 update.

[robin-raymond]

Ok, uppercase it will be with : between bytes.