Media/WebRTC/2014-03-25
Tues, March 25 at 10am Pacific.
Agenda/Notes:
1) Stand-ups:
- Active developers will give a 2 minute (or less) stand up focusing on -
- what are you currently working on that the team is following/tracking (please give the bug number) and when do you think you'll be done?
- what are you moving onto next?
- have you hit any problems/blockers (especially unexpected problems)?
- feel free to put a longer summary in the etherpad, but try to limit spoken comments to 2 mins or less. We can use the second half of the meeting (Discussion) for deeper dives.
- jesup -
- AEC patches up for review. Works well in limited testing, but I'll need to disable AEC in GUM when not connected to a peerconnection (unless overridden by a to-be-added constraint).
- Working on importing webrtc.org 3.50 and training pkerr on how to do it (and documenting it!)
- going to push to land the Talos audio quality patches
- Working on report of leaking threads (temporarily?)
- padenot - conflicting irl stuff, not attending
- reviewing jesup's patches for AEC in gUM
- finished documenting, and making green the MSG mixer patches (green as we speak)
- cubeb tests (now on all platforms)
- ekr - On travel, in meetings today
- abr -
- Largely in Loop design and discussion meetings
- Will work on the privileged vs web content issue ("chrome content issue") this week
- mt -
- Blocked on releng
- Got a few patches going out today
- ehugg - Conflicting meeting today.
- Debugging the GMP build of FF with OpenH264 - diagnosing crashes such as 987380 and updating it to build on Linux.
- jib -
- IdentityProvider reviews for Martin
- posted "Constraints 2014" proposal to public-media-capture,
- http://lists.w3.org/Archives/Public/public-media-capture/2014Mar/0083.html
- just finished slides for Thursday's w3c teleconference.
- bwc -
- Working on a huge review of PCP patches
- gcp -
- Opus ASM optimizations
- Attempts at profiling the above
- Should we start looking at that (profiling) more seriously?
- VP8 HW accel reviews
- H264 HW accel/Firefox OS
- drno (Nils) -
- buying dedicated hardware for the WebRTC QA tiger team
- hopefully close to solve the unexpected callback with message = "undefined" - bug 985274
- AEC kills one of the two fake audio signals we use in our mochitests
- dmose - mostly call standup stuff
- worked with Niko on landing repo layout for standalone client
- MLP UX review
- out part of last week for family issues
- code reviews
- up next: continue working with Std8 and Niko to get call up
- standard8 -
- Did some work on handling call information from the desktop conversation window
- Started integrating with SDKs to get the desktop conversation going (needs debugging)
- PTO on Thursday
- tOkeshu -
- Was on PTO thursday and friday
- paired with Niko yesterday
- worked on some bit server side and client side
- niko -
- mostly worked on the Loop webapp (for link clickers) https://bugzilla.mozilla.org/show_bug.cgi?id=979880
- had a successful call with tOkeshu yesterday using the client webapp & a fake page for the other end; synchronized with Mark to glue all the pieces together (the Loop desktop client, the server and the webapp) \o/
- alexis -
- deployed webapp and server
- writing documentation atm
- minor changes on the APIs
- implemented CORS on the server (webapp to be hosted on a CDN)
- added a way to discard calls
- loop server almost done (at least for MLP scope)
- natim / remy
- call status (to see if the caller rejected the call)
- working on a memcached backend (to expire some data)
- pkerr -
- completed re-work on bug768968 patches
- setting up to test load adaptation on OSX and Win7
- working through WebRTC 3.50 import and merge process that Randell has me checking out.
- florian - not attending; ping on IRC if questions
- I'm now confident the browser_devices_get_user_media.js tests will keep running. The intermittent failures (bug 976544) are mostly resolved:
- It stopped failing on mac after I renamed to test to get it to run before the fullscreen test that is suspected of messing with focus.
- I disabled the test on Linux Debug; but it's still running on Linux opt.
- Filed bug 983158 for a security review of persistent gUM permissions. Waiting for input from security.
- I'm now confident the browser_devices_get_user_media.js tests will keep running. The intermittent failures (bug 976544) are mostly resolved:
2) Discussion: a) questions from stand ups
- none
b) Platform quality for Loop(Bug 970426)
- Goal is to get all the essentials (known problems and critical tests) by Gecko 32.
- Landing talos tests and adding to them will be important, also telemetry
- I want to update from upstream webrtc.org soon, also upstream patches though that's lower priority
- Need to monitor upstream patches/bugs more closely.
- A few media element bugs we need to get addressed (video size in talky.io, loadedmetadata width/height, etc)
- Tuning gcp's load/bandwidth adaptation patch (important)
c) Orange Factor: tons of oranges for the unexpected-callback with "undefined" -- all appear to be in b2g-emulator runs d) MLP/MVP Loop
- Accountless mode (FxA signup not required) WILL be supported in MLP and MVP (though it is subject to user verification for MVP)
- Chrome content issue (Abr)j
- MVP doc (small edits still being made): https://docs.google.com/a/mozilla.com/spreadsheet/ccc?key=0AlJmiyFngeSwdGJEVnItQ2NVbGx6NXdjbVBUYmIxSGc#gid=0
e) Tuesday meeting
- discuss the best use of everyone's time -- after standups, it's up to each person individually to decide if they want to stay for the Discussion part of the meeting.
<Read only section>
Notices:
- Reminder to dogfood whenever possible: Use WebRTC for chats within the team whenever possible (prefer to IRC, etc if possible)
- Reviews:
- Reminder to handle review requests in a timely manner (previously discussed) and to raise any design issues early if possible, preferably outside of the the review process (in the bug, email, IRC, calls, whatever makes sense)
- Next meeting with tokbox will be today, April 1 at 2pm, Pacific.
References:
WebRTC Uplift Bugs: https://bugzilla.mozilla.org/buglist.cgi?list_id=6242297&resolution=FIXED&status_whiteboard_type=allwordssubstr&query_format=advanced&status_whiteboard=webrtc-uplift
Intermittent failures in WebRTC Automation: https://bugzilla.mozilla.org/buglist.cgi?keywords=intermittent-failure%2C%20&keywords_type=allwords&list_id=6167884&resolution=---&query_format=advanced&component=WebRTC&component=WebRTC%3A%20Audio%2FVideo&component=WebRTC%3A%20Networking&component=WebRTC%3A%20Signaling&product=Core