Closed Bug 1094447 Opened 10 years ago Closed 9 years ago

Start using UDP/TLS/RTP/SAVPF in offers once it interops

Categories

(Core :: WebRTC: Signaling, defect, P3)

defect

Tracking

()

RESOLVED FIXED
mozilla43
Tracking Status
firefox43 --- fixed

People

(Reporter: bwc, Assigned: bwc)

References

Details

Attachments

(1 file)

Current specs say we should use UDP/TLS/RTP/SAVPF instead of RTP/SAVPF, but this does not play well with currently deployed firefox, and possibly other implementations. Bug 1091242 will allow us to accept this in an offer, and once this work is on all supported branches (or stopgaps are put in place), we should investigate making this change.
backlog: --- → webRTC+
Rank: 50
Priority: -- → P4
Looks like the Chrome guys have committed the same fix on their end:

https://chromium.googlesource.com/external/webrtc.git/+/f39382943449b7e44ac563e05a14203534591acf

We should go ahead and do this now.
Assignee: nobody → docfaraday
Rank: 50 → 30
Priority: P4 → P3
Bug 1094447: Use UDP/TLS/RTP/SAVPF for audio/video m-lines
Comment on attachment 8655682 [details]
MozReview Request: Bug 1094447: Use UDP/TLS/RTP/SAVPF for audio/video m-lines

https://reviewboard.mozilla.org/r/17977/#review16111

::: media/webrtc/signaling/test/signaling_unittests.cpp:3390
(Diff revision 1)
>    // TODO(bug 1099351): Use commented out code instead.

Can we update this comment to point to bug 814227 please?
Attachment #8655682 - Flags: review+
Comment on attachment 8655682 [details]
MozReview Request: Bug 1094447: Use UDP/TLS/RTP/SAVPF for audio/video m-lines

https://reviewboard.mozilla.org/r/17977/#review16115

::: media/webrtc/signaling/test/signaling_unittests.cpp:3240
(Diff revision 1)
> -    "m=audio 56187 RTP/SAVPF 109 0 8 101\r\n"
> +    "m=audio 56187 UDP/TLS/RTP/SAVPF 109 0 8 101\r\n"

We probably want at least one test case which keeps using just 'RTP/SAVPF' to ensure we don;t loos the ability to handle that properly and answer with 'RTP/SAVPF' in that case.
Attachment #8655682 - Flags: review+
https://reviewboard.mozilla.org/r/17977/#review16115

> We probably want at least one test case which keeps using just 'RTP/SAVPF' to ensure we don;t loos the ability to handle that properly and answer with 'RTP/SAVPF' in that case.

I can add such a test to jsep_session_unittest
Comment on attachment 8655682 [details]
MozReview Request: Bug 1094447: Use UDP/TLS/RTP/SAVPF for audio/video m-lines

Bug 1094447: Use UDP/TLS/RTP/SAVPF for audio/video m-lines
Attachment #8655682 - Flags: review?(drno)
Check back on try.
Flags: needinfo?(docfaraday)
Flags: needinfo?(docfaraday)
Attachment #8655682 - Flags: review?(drno) → review+
Comment on attachment 8655682 [details]
MozReview Request: Bug 1094447: Use UDP/TLS/RTP/SAVPF for audio/video m-lines

https://reviewboard.mozilla.org/r/17977/#review16207

We could go as crazy as testing for switching protocols mid-flight. But I don't think we actually want to support that.
Keywords: checkin-needed
Status: NEW → RESOLVED
Closed: 9 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla43
I've already sent mail to mozilla-dev-media about this, and we need to make sure this is mentioned in the release notes as well.
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: