mirror of https://github.com/OpenVidu/openvidu.git
23 tests for a participant publishing more than one audio or video track: screen share and its audio next to the camera and the microphone, before them, alone, muted, stopped and resumed, two cameras, two microphones, the first of each unpublished, six tracks, mixed codecs, screen_share mode, MediaSource with two audios, WebM, publisher packet loss with two video tracks, and the manifest's track list against ffprobe. The synthetic screen gets an audio track when the capture asks for one, the Publisher helper drives the screen options dialog and addresses cameras and microphones by track sid, since the testapp lists tracks by sid, and parks the mouse after the add-track buttons. Also publisherPacketLossH264Test, publisherPacketLossAv1Test, publisherJitterVp9Test and publisherJitterAv1Test. The AV1 loss test found the SFU switching temporal layers up on a frame whose T1 reference it had dropped (openvidu-pro-mediasoup 4326f0c19); it carries a TODO and fails on mediasoup until the server image ships a worker released from that commit. Assisted-by: claude-code - Anthropic - Claude Fable 5.1 |
||
|---|---|---|
| .. | ||
| docker | ||
| src | ||
| pom.xml | ||