Add negative tests for prefixed variants of MediaStream, Speech and WebRTC #7507
Conversation
wpt-pr-bot
requested review from
agouaillard,
alvestrand,
dontcallmedom,
guidou,
henbos,
phoglund,
rwaldron and
youennf
|
@lukebjerring, this is a typical case where #7475 would really increase confidence in the changes. I've written the tests by copy-pasting the API names from various places, but any typo would cause the tests to pass. Being able to easily see what the pass/fail status of all new tests are would make it clear at a glance if I got it right, since I know which of these are in Chrome (all of them). @drufball, if you're doing a PRD for this work, maybe this can be one of the use cases. The other big one is of course when changing tests. |
|
@andrenatal @gshires, I added you as OWNERS of speech-api since this is the first test there. Are you OK with that? |
Build PASSEDStarted: 2017-09-27 21:54:55 View more information about this build on: |
Tests should follow what the specs say, and unless the specs define the prefixed APIs [1] the right this is to assume the per-spec names and to add historical.html tests asserting the absense of prefixed APIs that have existed somewhere. Based on web-confluence.appspot.com, navigator.vibrate was never prefixed, but mozSrcObject is in Firefox. A test is added for mozSrcObject, and the others APIs are already covered by existing tests, or one of these: #7507 #7508 [1] https://dom.spec.whatwg.org/#dom-element-webkitmatchesselector https://url.spec.whatwg.org/#url-class and others
webrtc/historical.html
Outdated
|
|
||
| test(function() { | ||
| assert_false("webkitRTCPeerConnection" in window); | ||
| }, "webkitRTCPeerConnection interface should not exist"); |
There was a problem hiding this comment.
Since you've add mozGetUserMedia in mediacapture-streams/historical.html, should this file also include:
mozRTCIceCandidate
mozRTCPeerConnection
mozRTCSessionDescription
?
There was a problem hiding this comment.
Yes, for sure! I didn't know they existed :)
|
@fleizach, added you here, now maybe you can review? :) |
Tests should follow what the specs say, and unless the specs define the prefixed APIs [1] the right this is to assume the per-spec names and to add historical.html tests asserting the absense of prefixed APIs that have existed somewhere. Based on web-confluence.appspot.com, navigator.vibrate was never prefixed, but mozSrcObject is in Firefox. A test is added for mozSrcObject, and the others APIs are already covered by existing tests, or one of these: #7507 #7508 [1] https://dom.spec.whatwg.org/#dom-element-webkitmatchesselector https://url.spec.whatwg.org/#url-class and others
|
@bobholt, when I was showing this to @drufball I noticed that https://pulls.web-platform-tests.org/build/18921 didn't have results for some of the browsers, I think it said "No tests run" or something. But now I'm looking again and they're there. This was all today, well after "Build PASSED" was posted. Any idea what might be going on? |
|
I think you must have clicked through to the test-by-test results. It looks
like they're not showing up for Firefox or Chrome (
https://pulls.web-platform-tests.org/job/18921.10), but are showing up for
Safari and Edge (https://pulls.web-platform-tests.org/job/18921.11).
I'll take a look.
…On Fri, Sep 29, 2017 at 2:31 PM, Philip Jägenstedt ***@***.*** > wrote:
@bobholt <https://github.com/bobholt>, when I was showing this to
@drufball <https://github.com/drufball> I noticed that
https://pulls.web-platform-tests.org/build/18921 didn't have results for
some of the browsers, I think it said "No tests run" or something. But now
I'm looking again and they're there. This was all today, well after "Build
PASSED" was posted. Any idea what might be going on?
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#7507 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AAFv10rtnW9y4GPw3pG0JHlWK7z1w8hiks5snTeDgaJpZM4PlvQO>
.
|
Tests should follow what the specs say, and unless the specs define the prefixed APIs [1] the right this is to assume the per-spec names and to add historical.html tests asserting the absense of prefixed APIs that have existed somewhere. Based on web-confluence.appspot.com, navigator.vibrate was never prefixed, but mozSrcObject is in Firefox. A test is added for mozSrcObject, and the others APIs are already covered by existing tests, or one of these: web-platform-tests#7507 web-platform-tests#7508 [1] https://dom.spec.whatwg.org/#dom-element-webkitmatchesselector https://url.spec.whatwg.org/#url-class and others
No description provided.