Closed Bug 1302621 Opened 3 years ago Closed 3 years ago
_context _name .html | Retaining window .name on history traversal - assert _equals: Initial load expected "test" but got ""
James and/or Ms2ger, the timing of this appears to coincide with the upstream update from bug 1302423. It appears there was a change to that file included in the update, though it seems fairly innocuous? https://github.com/w3c/web-platform-tests/commit/81207ecf9020aaea95cda2170a17421fd9d3f660 Anyway, this is a pretty regular failure on Windows mostly (but sometimes other platforms as well), so can you please take a look?
Component: web-platform-tests → Document Navigation
Product: Testing → Core
See Also: → 1308969
Version: Version 3 → Trunk
It looks like the problem is here: onload = setTimeout(next, 0); Possibly we end up running the first step as a result of the load event dispatched at the iframe, and then run the second step from the timeout, at which point the name is not yet set. I'm not sure how my change would have affected that, though...
Flags: needinfo?(Ms2ger) → needinfo?(geoffers+mozilla)
I have no idea about these tests; I may have originally committed them to wpt but they were written by James.
Comment on attachment 8828801 [details] Bug 1302621 - Fix buggy history traversal test, https://reviewboard.mozilla.org/r/106096/#review107860
Attachment #8828801 - Flags: review?(Ms2ger) → review+
Pushed by email@example.com: https://hg.mozilla.org/integration/autoland/rev/6f97023c9af9 Fix buggy history traversal test, r=Ms2ger
Pushed by firstname.lastname@example.org: https://hg.mozilla.org/integration/autoland/rev/6462429dd8eb fixup wpt manifest, a=testonly
James, would it be possible to whip up a Beta52 backport patch for this failure? It's pretty frequent on that branch and destined to be our next ESR. I took a quick stab at it but got caught in the weeds pretty quickly.
You need to log in before you can comment on or make changes to this bug.