Investigate failing webdriver tests from wpt downstream sync bug 1709044
Categories
(Testing :: geckodriver, defect, P2)
Tracking
(firefox89 unaffected, firefox90 affected)
Tracking | Status | |
---|---|---|
firefox89 | --- | unaffected |
firefox90 | --- | affected |
People
(Reporter: whimboo, Unassigned)
References
(Regression)
Details
(Keywords: regression)
A recent downstream sync for Remove use of six.integer_types/string_types/text_type
(bug 1709044) brought a lot of failing webdriver tests:
https://hg.mozilla.org/integration/autoland/rev/1b852017ee57
It's mostly related to user prompt tests and browsing contexts. Seeing the actual changes it's not clear to me why those tests fail now.
Getting lots of these meta data silently added, and not pro-actively checking the wpt sync bugs in detail for those, a test suite will not be trustworthy at some time. Lets discuss in our next triage meeting how we can better handle that.
Reporter | ||
Updated•3 years ago
|
Reporter | ||
Comment 1•3 years ago
|
||
Actually not sure what happened here. While the above mentioned revision shows the file as being present, I do not see it anymore in the latest revision:
Also hg log
doesn't show anything that this file got removed. James, what's going on here?
Updated•3 years ago
|
Comment 2•3 years ago
|
||
Lookd like the try run had some new failures, but they disappeared in the landing commit. Could mean that the failures went away with a subsequent commit or that the try run was outdated/misleading for some reason (there was definiely some churn with the PR with earlier revisions causing regressions).
Reporter | ||
Comment 3•3 years ago
|
||
Going to mark as WFM given that no new multiple status changes have finally been landed.
Updated•3 years ago
|
Description
•