Our hold implementation has historically been flaky - requiring application level "workarounds" to make it work reliably. We have also depended rewriting the SDP to make it work, which is less than ideal (RFC8829 makes it clear that is not an approach which will work reliably).

• Fixes SimpleUser mute & hold issues
• Fixes Web SDH issue un-holding when peers put each other on hold
• RFC8829 (Jan 2021) compliant implementation (no longer uses "hold modifier" to rewrite SDP)
• Added demo to see all possible mute and hold combinations function
• Added tests for Web SDH hold

Note that our Web SDH hold implementation now depends on RTCRtpTransceiver support being available. While the implementation guards against it not being present, SimpleUser.hold will not work without it (it will noop). That said the prior implementation didn't actually work in all cases, so...