Start running mochitest-a11y with e10s enabled in automation

RESOLVED WONTFIX

Status

Release Engineering
General
RESOLVED WONTFIX
2 years ago
12 days ago

People

(Reporter: RyanVM, Unassigned)

Tracking

(Blocks: 1 bug)

Firefox Tracking Flags

(Not tracked)

Details

(Whiteboard: aes+)

(Reporter)

Description

2 years ago
Now that Aaron is getting close to getting e10s a11y support ready for widespread testing, we need to get automated tests running in automation.

Explicit non-goal for this bug: splitting up non-e10s mochitest-other across all platforms. We should probably consider doing that soon to avoid confusion, but we can take care of this bug without having to do that.
(Reporter)

Comment 1

2 years ago
To the people CCed: feel free to add any other deps you want.
(Reporter)

Updated

2 years ago
Blocks: 1186029
Depends on: 1304449
(Reporter)

Comment 2

2 years ago
Getting this running on Taskcluster is apparently trivial. So that's awesome.
https://treeherder.mozilla.org/#/jobs?repo=try&revision=2e09e0469bf62bafe4b72e7bef3f1c6092072ff7
except that is linux only- but a start!
(Reporter)

Comment 4

2 years ago
Yeah, good enough for ensuring that they're not completely and utterly broken before I embark on the buildbot end of things.

Updated

2 years ago
Blocks: 984139
Ryan, is there anything for my team to do here?
Flags: needinfo?(ryanvm)
(Reporter)

Comment 6

2 years ago
This got held up a bit by questions about how useful M-e10s(a11y) would actually be. I'll try to work on it soon, though.
Flags: needinfo?(ryanvm)
(Reporter)

Comment 7

2 years ago
Finally got Linux working with some hacks. And they're currently busted.
https://treeherder.mozilla.org/logviewer.html#?job_id=29806780&repo=try

Though I'm still a bit confused about what I'm seeing in the logs. As expected, the beginning of the run shows "runtests.py | Running with e10s: True" in the logs, but at the end of the run, I still see "INFO Mode: non-e10s". Not entirely sure what to make of that :\
(Reporter)

Comment 8

2 years ago
After a number of discussions with various accessibility people, the unanimous conclusion is that there isn't any value in trying to stand up M-e10s(a11y). It won't effectively test cross-process things anyway and time is better-spent on adding more browser-chrome tests to the ones that are already there.
Status: NEW → RESOLVED
Last Resolved: 2 years ago
Resolution: --- → WONTFIX
(Assignee)

Updated

12 days ago
Component: General Automation → General
Product: Release Engineering → Release Engineering
You need to log in before you can comment on or make changes to this bug.