W3C home > Mailing lists > Public > public-webrtc-logs@w3.org > January 2018

Re: [webrtc-pc] RTCRtpContributingSource.audioLevel has different type and range than similar fields in webrtc-stats

From: Philip Jägenstedt via GitHub <sysbot+gh@w3.org>
Date: Fri, 19 Jan 2018 16:48:33 +0000
To: public-webrtc-logs@w3.org
Message-ID: <issue_comment.created-359023684-1516380512-sysbot+gh@w3.org>
Hi all,

@rakuco and I are looking at [normative spec change without tests](https://docs.google.com/document/d/1kFYTqUMEbo9p87i6wAzORoicMWgHsuYwL2CgtvdhDaE/edit?usp=sharing) every week this quarter, and this week I took a look at https://github.com/w3c/webrtc-pc/pull/1736, since I'm a bit familiar with the tests that were written.

There are tests for `audioLevel` being a byte written by @soareschen (reviewed by @rwaldron) that now contradict what the spec says, in [RTCRtpReceiver-getSynchronizationSources.https.html](https://github.com/w3c/web-platform-tests/blob/524bea1dd8d0ad58733072bd4e8b8044042f589e/webrtc/RTCRtpReceiver-getSynchronizationSources.https.html#L64) and [RTCRtpReceiver-getContributingSources.https.html](https://github.com/w3c/web-platform-tests/blob/524bea1dd8d0ad58733072bd4e8b8044042f589e/webrtc/RTCRtpReceiver-getContributingSources.https.html#L77). (Found by grep.)

I've filed https://github.com/w3c/web-platform-tests/issues/9108 to keep track of the fact that the tests are now wrong, and looking through https://github.com/w3c/web-platform-tests/labels/webrtc before process transitions would probably be a good idea since the test results could otherwise be misleading, *especially* if someone has implemented the old behavior and passes those tests.

[CONTRIBUTING.md](https://github.com/w3c/webrtc-pc/blob/master/CONTRIBUTING.md#tests) suggests at least filing an issue like that, but this is a case where I think blocking the spec change on the test update would have been less work overall.

But I'm not just here to nag, if there are any bits of infrastructure that you all think would be useful to make it harder to overlook testing, I'd love to hear some ideas.

https://github.com/whatwg/html/issues/1849#issuecomment-351359904 describes the current state of this "project", and of particular interest could be the auto-filing of bugs, which in the best case can mean that if Chrome was passing a test and test was changed to fail, a bug would be automatically filed.

GitHub Notification of comment by foolip
Please view or discuss this issue at https://github.com/w3c/webrtc-pc/issues/1734#issuecomment-359023684 using your GitHub account
Received on Friday, 19 January 2018 16:48:35 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 19:21:58 UTC