1903663 - Perma failure /css/cssom/CSSStyleSheet-constructable-concat.html | single tracking bug

Status: RESOLVED FIXED

(Core :: DOM: CSS Object Model, defect)

Description

[Treeherder Bug Filer]

Filed by: istorozhko [at] mozilla.com
Parsed log: https://treeherder.mozilla.org/logviewer?job_id=462066979&repo=try
Full log: https://firefox-ci-tc.services.mozilla.com/api/queue/v1/task/ErhPrR4QR5ioveD_trSWwg/runs/0/artifacts/public/logs/live_backing.log
Reftest URL: https://hg.mozilla.org/mozilla-central/raw-file/tip/layout/tools/reftest/reftest-analyzer.xhtml#logurl=https://firefox-ci-tc.services.mozilla.com/api/queue/v1/task/ErhPrR4QR5ioveD_trSWwg/runs/0/artifacts/public/logs/live_backing.log&only_show_unexpected=1

---

This test fails on all Android Fission builds,
with both `isolate-everything` and `isolate-high-value` Fission site isolation strategies.

Blocks Fission on Android

Comment 1

[[:owlish] 🦉 PST]

(I used the intermittent failure filer on Treeherder, but this is a permanent failure on Android Fission)

Comment 2

[Daniel Holbert [:dholbert]]

It's actually TEST-UNEXPECTED-PASS, not a failure. :)

The test's only annotation is:
https://searchfox.org/mozilla-central/rev/e729cdd1537458f9268215c2ff6dddb231c3a6d9/testing/web-platform/meta/css/cssom/CSSStyleSheet-constructable-concat.html.ini#1-3

[CSSStyleSheet-constructable-concat.html]
  expected:
    if (os == "android") and fission: [OK, TIMEOUT]

which looks like it's intending to allow the test to either pass or timeout.

I suspect the issue is the OK status there -- I think we use that for testharness.js-based tests, but for reftest-style tests (which this one is), we need PASS instead.

owlish, would you mind testing s/OK/PASS/ in that^ .ini file and see if that fixes the issue? Alternately: if this really does always pass, we could just remove the .ini file entirely, since it's expressing an allowance that we perhaps don't need anymore (and doing so in a way that's not-quite-right, per this bug).

EDIT: it's probably not worth testing one-off tweaks here, per my next comment.)

Comment 3

[Daniel Holbert [:dholbert]]

(see also bug 1903914 comment 3 - 4 for more thoughts; this bug might really want to be duped to the bug that's already removing the failure annotation, bug 1522790)

Comment 4

[[:owlish] 🦉 PST]

Hi Daniel, I replied in bug 1903914. Basically, my question is, if we set the current annotations aside for a second, does this test actually supposed to pass or fail? (if it's supposed to pass, then of course, I'll switch the annotation, and the problem will be solved. But if it's supposed to fail..... Then we are having incorrect behaviour with Fission?)

Comment 5

[Daniel Holbert [:dholbert]]

See my answer in bug 1903914 comment 4; it applies to all of these bugs. (There's no reason for there to be an android-with-fission specific expectation in any of these, aside from the fact that it was previously broken and we were annotating that known-breakage.)

Comment 6

[[:owlish] 🦉 PST]

Attachment: Bug 1903663 - Adjust test expectation for Android Fission builds

Comment 7

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/5daf2b680bae
Adjust test expectation for Android Fission builds r=dholbert

Comment 8

[Iulian Moraru]

https://hg.mozilla.org/mozilla-central/rev/5daf2b680bae