Media/WebRTC/2012-04-19

From MozillaWiki
< Media‎ | WebRTC
Jump to: navigation, search

Participants:

Maire Reavy (Mozilla), Randell Jesup (Mozilla), Ethan Hunt (Cisco), Enda Mannion (Cisco), Suhas Nandaku (Cisco), Eric Rescorla (nick:ekr, Mozilla), Cullen Jennings (Cisco), Tim Terriberry (Mozilla)

Notes:

1) Discussed the state of the paris_demo branch:

  • How close are we to getting Xavier (from Spacegoo) unblocked?
    • He needs patches from bug 744933 and 745259 since he runs Ubuntu - Both patches just pushed to paris_demo.
    • And copy a .yuv into camera.yuv in the default directory if dual-instancing.
    • We identified the Alsa bug and filed a webrtc.org issue on it; patch RSN

2) Discussed Enda's doc (a simple header file on top of the sipcc engine, that defines the interface that Randell can call into).

  • It would be very useful to have a description of how that interfaces with sipcc to go along with the header file.
  • Enda will take EKR's changes and incorporate them into his code within the next day.
  • Need ICE ufrag/fingerprints/etc and candidates in the SDP
  • Need to refactor SIPCC interface to be driven from above; design a base class to specify the mediastreams and constraints/hints needed to generate the SDP
  • Note that encoded video streams may be sources, and may need to be transcoded if the other side doesn't support it, or if we need to scale, crop, etc, etc. Currently the webrtc.org code does support encoded sources.
  • State will need to live in one of these: PeerConnection, with SIPCC largely stateless; SIPCC, with PeerConnection/etc largely stateless, or both and make sure they stay in sync.
  • Enda will refactor and then go over it with Eric on Monday

3) Discussed need for estimate of when we think we'll have basic proper signaling working -- i.e. video/audio being sent and received so that we can say we have a basic call on a Windows machine. (We're not interested in a demo. We're interested in a test rig.)

  • Once we have sipcc refactored, we should have working calls on all platforms (Windows, Mac, Linux) on Alder trunk. At that point, we can try making WebRTC calls to Chrome. We should have an estimate next week once we pull the schedule together.

4) Reviewed project map

  • Some items are "rollup" nodes - i.e. "Refactor SIPCC" is really "SIPCC Refactored"
  • Landing plans for m-c are largely being kept out of this graph - separate plan/schedule (though dependent on this one).
  • Dates to be left on for now; after we know we have a good graph we'll need to develop real dates and update with them. (Largely next week)

5) Discussed plan for initial landings

  • MediaStreams will be landing in m-c soon (maybe FF14)
  • Initial landing of code on m-c can proceed without sec review if it's preffed off.
  • Randell will coordinate with build and other peers before landing webrtc.org code
  • Support for getUserMedia() (basic support) is likely to land first and enables a lot of testing; also likely to be the first parts reviewed and preffed on.

6) No new items mentioned

7) Discussed upcoming goals and when we need them:

  • Get updates on the dependency graph and any estimates you have by mid-day monday (noon PDT). Maire will be working with Randell, Ethan, EKR on Monday afternoon to assign names to tasks before our Tuesday morning weekly meeting.

8) Changing weekly meeting time: proposal from EKR-->8:30 AM

  • Decision on the call for new date/time: Tuesdays, 8:30AM PDT, starting May 1st
  • Enda says he can be flexible on weekly meeting day if we want a later meeting, like 11,12, 1 PST, an earlier time as EKR proposes is perfect for him also.