1539693 - windows/aarch64 - re-enable/adjust web-platform-tests results based on new timeout multiplier

Status: RESOLVED FIXED

(Testing :: web-platform-tests, enhancement)

Description

[Edwin Takahashi (:egao | infrequent contributor)]

In bug 1533880 the ongoing task was to disable many of the web-platform-tests due to various unexpected outcomes for windows10-aarch64.

Since bug 1538725 landed it may be possible to re-enable many of the tests that were simply disabled due to TEST-UNEXPECTED-TIMEOUT or TEST-UNEXPECTED-FAIL. Other tests that used to fail, or timeout are now passing and require its outcome be adjusted.

Outcome adjustment:

• /html/infrastructure/urls/resolving-urls/query-encoding/navigation.sub.html?encoding=windows-1252
• /html/infrastructure/urls/resolving-urls/query-encoding/navigation.sub.html?encoding=utf8
• /html/infrastructure/urls/resolving-urls/query-encoding/navigation.sub.html?encoding=x-cp1251

Re-enable:

• /webrtc/RTCPeerConnection-getStats.https.html

Comment 1

[Edwin Takahashi (:egao | infrequent contributor)]

Attachment: Bug 1539693 - adjust expected outcomes and/or re-enable tests in wpt for windows10-aarch64

Comment 2

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/47decffef055
adjust expected outcomes and/or re-enable tests in wpt for windows10-aarch64 r=jmaher

Comment 3

[Andrei Ciure[:aciure]]

https://hg.mozilla.org/mozilla-central/rev/47decffef055

Comment 4

[Byron Campen [:bwc]]

I see that testing/web-platform/meta/webrtc/RTCDataChannel-send.html.ini and testing/web-platform/meta/webrtc/RTCRtpTransceiver.https.html.ini still disable their tests on aarch64; are those tests not working with the new timeout multiplier?

Comment 5

[Edwin Takahashi (:egao | infrequent contributor)]

(In reply to Byron Campen [:bwc] from comment #4)

I see that testing/web-platform/meta/webrtc/RTCDataChannel-send.html.ini and testing/web-platform/meta/webrtc/RTCRtpTransceiver.https.html.ini still disable their tests on aarch64; are those tests not working with the new timeout multiplier?

I think that was simply an oversight. I've pushed to try with the tests enabled, see https://treeherder.mozilla.org/#/jobs?repo=try&resultStatus=pending%2Crunning%2Csuperseded%2Cusercancel%2Cretry%2Csuccess%2Ctestfailed%2Cbusted%2Cexception&classifiedState=unclassified&group_state=expanded&revision=49f08b89a0799bdeae6a209b533ffcd98dd70152.

Comment 6

[Edwin Takahashi (:egao | infrequent contributor)]

According to the try push, it seems these tests are still timing out or not running. It looks better to remain disabled, since I recall the outcomes are intermittent and change when expected outcomes are defined for some tests.

Comment 7

[Byron Campen [:bwc]]

Yeah, I'm working on those over in bug 1542343 and bug 1542345.