Closed
Bug 1330091
Opened 8 years ago
Closed 8 years ago
Renegotiation doesn't actually change the codec configuration after 49 update landing
Categories
(Core :: WebRTC, defect, P1)
Core
WebRTC
Tracking
()
RESOLVED
FIXED
mozilla53
Tracking | Status | |
---|---|---|
firefox50 | --- | unaffected |
firefox51 | --- | unaffected |
firefox52 | --- | unaffected |
firefox53 | --- | fixed |
People
(Reporter: jesup, Assigned: jesup)
References
Details
(Keywords: regression)
Attachments
(2 files)
2.48 KB,
patch
|
ng
:
review+
|
Details | Diff | Splinter Review |
3.34 KB,
patch
|
ng
:
review+
|
Details | Diff | Splinter Review |
The rewrite of ConfigureSendMediaCodec() ends up not actually installing the new codec/settings it calculates, since StartTransmitting() checks mSendStream, and if it's set doesn't instantiate a new stream (and encoder). In fact, we want to ReconfigureVideoCodec() where at all possible.
Assignee | ||
Updated•8 years ago
|
Rank: 13
status-firefox53:
--- → affected
Assignee | ||
Comment 1•8 years ago
|
||
Also fixes that on a renegotiation it didn't install the new config at all
Attachment #8825533 -
Flags: review?(na-g)
Comment 2•8 years ago
|
||
Comment on attachment 8825533 [details] [diff] [review]
Reconfigure video encoders where possible instead of recreating them
Review of attachment 8825533 [details] [diff] [review]:
-----------------------------------------------------------------
One potential issue.
::: media/webrtc/signaling/src/media-conduit/VideoConduit.cpp
@@ +466,5 @@
> if ((condError = ValidateCodecConfig(codecConfig, true)) != kMediaConduitNoError) {
> return condError;
> }
>
> + if (!mSendStream ||
Should this be if(mSendStream && ... ?
Attachment #8825533 -
Flags: review?(na-g) → review-
Comment 3•8 years ago
|
||
Comment on attachment 8825533 [details] [diff] [review]
Reconfigure video encoders where possible instead of recreating them
Review of attachment 8825533 [details] [diff] [review]:
-----------------------------------------------------------------
Looks good.
Attachment #8825533 -
Flags: review- → review+
Comment 4•8 years ago
|
||
Can you add a mochitest that negotiates with a non-default codec and verifies that content is making it across?
Probably one explicit test per supported codec.
And then one test that renegotiates and changes codecs to test this bug.
It's a bit of work but since they're tests they'll save us work in the long run! :-)
Updated•8 years ago
|
Assignee | ||
Comment 5•8 years ago
|
||
Attachment #8825922 -
Flags: review?(na-g)
Comment 6•8 years ago
|
||
Comment on attachment 8825922 [details] [diff] [review]
interdiff for h264 compatibility checking, plus a bugfix
Review of attachment 8825922 [details] [diff] [review]:
-----------------------------------------------------------------
r+ on the interdiff, with a question. Does the h264 level also need to be checked?
Also, +1 to adding mochitest coverage.
Attachment #8825922 -
Flags: review?(na-g) → review+
Assignee | ||
Comment 7•8 years ago
|
||
"49" here means bug 1250356 (import webrtc.org stable branch 49) not Firefox 49
Assignee | ||
Comment 8•8 years ago
|
||
h.264 level is negotiated, and can change in an answer for example. Changing it would require a new SPS/PPS to be sent, much like a resolution change.
Pushed by rjesup@wgate.com:
https://hg.mozilla.org/integration/mozilla-inbound/rev/abf0cffb12f7
Reconfigure video encoders where possible instead of recreating them r=ng
Comment 10•8 years ago
|
||
bugherder |
Status: NEW → RESOLVED
Closed: 8 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla53
Updated•8 years ago
|
status-firefox50:
--- → unaffected
status-firefox51:
--- → unaffected
status-firefox52:
--- → unaffected
Version: 45 Branch → Trunk
You need to log in
before you can comment on or make changes to this bug.
Description
•