Handling of duplicate rids in sRD seems underspecified

[docfaraday]

For example, a=simulcast:recv foo;foo

While this is pretty silly, I don't see anything in RFC 8853 that forbids it. I see two reasonable options:

1. Ignore any duplicates.
2. Fail the sRD, probably with an InvalidAccessError.

I think I prefer option 1. We might want to consider consistency with #2769

[jan-ivar]

Ignore any duplicates.

Do you mean A) ignore the second foo, or B) ignore both foos?

[docfaraday]

I mean A.

[jan-ivar]

It seems simple enough to mandate that an answer ignore duplicates in an offer.

[dontcallmedom-bot]

This issue was discussed in WebRTC January 2023 meeting – 17 January 2023 (Issue #2780 duplicate rids in sRD underspecified 🎞︎)