Conversation
reillyeon
left a comment
There was a problem hiding this comment.
Test changes that I would probably be considered an owner of look good to me. Didn't look at any of the automation infrastructure changes.
595e425 to
c5aa0e5
Compare
infrastructure/metadata/infrastructure/testdriver/set_permission.https.html.ini
Show resolved
Hide resolved
rakuco
left a comment
There was a problem hiding this comment.
lgtm
The Chromium failure in the checks looks unrelated; I'd try pushing again to see if it goes away, otherwise someone with enough power will have to force-merge this PR.
c5aa0e5 to
92772bf
Compare
|
@rakuco indeed. @jgraham maybe you can override it? @reillyeon, as spec editor, any chance you could check if idle-detection/idle-detection-allowed-by-permissions-policy.https.sub.html is indeed flaky? It does to a lot of cross origin stuff, so it certainly has the potential to be flaky. |
| @@ -1,17 +1,23 @@ | |||
| <!DOCTYPE html> | |||
| <meta charset="utf-8"> | |||
| <meta charset="utf-8" /> | |||
There was a problem hiding this comment.
No, ran prettier over it… it adds the “ />” garbage.
| await test_driver.set_permission(descriptor, "denied"); | ||
| permission = await navigator.permissions.query(descriptor); | ||
| assert_equals(permission.state, "denied"); | ||
| }, "Deny Permission"); |
There was a problem hiding this comment.
Should we really be changing this entire test as part of this?
There was a problem hiding this comment.
In this case, yes. It doesn’t do any harm to test that setting the permission worked.
The use of realm and "oneRealm" are being removed from the Permission spec.
w3c/permissions#397
And related discussion:
w3c/permissions#387