Additionally, we recently faced an issue and it might make sense to
introduce some changes in the revision for this.
Currently, section 7.5.3 says "If two m lines are grouped using FID they
MUST differ in their transport addresses (i.e., IP address plus port)."
While I believe we should have this requirement, sometimes it is
possible to have multiple m lines in an SDP that use the same transport
address, but only one of those streams is active at a time (others being
inactive).
Usually, an active stream wouldn't need to be grouped with inactive
stream(s). However, RFC 4566 says "The semantics of multiple "m=" lines
using the same transport address are undefined." And in practice, some
non-100%-compliant SDP parsers may reject the whole SDP simply if they
observe multiple m lines sharing the same transport address, without
checking whether these m lines were grouped or not.
Maybe, we could introduce an escape clause for inactive streams (against
being checked whether they use a transport address that is also used by
another m line). I am not sure whether anybody else has a comment on
this, but if there is a need, we might fix it in this revision.