Closed Bug 809233 Opened 7 years ago Closed 7 years ago

Get B2G xpcshell tests running on cedar


(Testing :: XPCShell Harness, defect, P1)



(blocking-basecamp:+, b2g18 fixed)

B2G C2 (20nov-10dec)
blocking-basecamp +
Tracking Status
b2g18 --- fixed


(Reporter: ahal, Assigned: ahal)




(1 file)

Now that mochitests and reftests are running, we're at the point where we want to get xpcshell tests running on cedar so we can get them checked in to mozilla central.
Depends on: 797153
So we'll need to update mozharness to run xpcshell tests and then enable it on B2G. However we aren't even at the stage where the harness is ready to run on B2G at all. This bug will track the progress towards that point.
I have been working on xpcshell-for-android lately; some of those changes might help xpcshell-for-b2g -- see bug 768491 and its dependents.
Depends on: 815641
Summary: Get xpcshell tests running on cedar → Get B2G xpcshell tests running on cedar
Blocks: 787578
Depends on: 815770
Blocks: 810547
Depends on: 816079
Depends on: 816086
Depends on: 790817
Depends on: 816612
No longer depends on: 816086
Depends on: 816630
Depends on: 817122
Depends on: 817604
Depends on: 818152
This is live on cedar now, but there are two issues that I didn't see when testing locally we'll have to fix.

1) busybox unzip is failing for some reason (bug 818152)
2) it looks like the full set of tests is being run instead of the xpcshell_b2g.ini one
2b) a side effect is that one of the tests that ends up being run times out and the job gets killed

I need to spend a bit of time debugging.
Depends on: 818259
Issue 1 is due to using something new in python 2.7 (the talos-r3-fed-xxx slaves have something older). Should be fixed soon.
Depends on: 818638
Depends on: 819399
blocking-basecamp: --- → +
Target Milestone: --- → B2G C2 (20nov-10dec)
Priority: -- → P1
Quick update: this is potentially ready to resolve now. Unfortunately wait times on cedar are taking almost an entire work day, so it's been a very slow triaging process. All that should be left though is to figure out if there are any other intermittent oranges, disable them on m-c, fix bug 819399 (trivial) and hand it off to releng to create/enable the proper configs.
Heads up: my patch in bug 807125 enables emulator xpcshell on all branches through b2g18, so we'll need to either hide it or uplift your fixes.

Let me know if we should limit that list for some reason.
Attachment #691139 - Flags: review?(ahalberstadt)
Comment on attachment 691139 [details] [diff] [review]
xpcshell exit code 1 -> INFO

Review of attachment 691139 [details] [diff] [review]:

Lgtm. Not sure what you mean by 'limit that list'. Are you talking about the manifest?
Attachment #691139 - Flags: review?(ahalberstadt) → review+

These are running green on cedar and are now ready to go on central, aurora and b2g18
Closed: 7 years ago
Resolution: --- → FIXED
(In reply to Andrew Halberstadt [:ahal] from comment #8)
> Comment on attachment 691139 [details] [diff] [review]
> xpcshell exit code 1 -> INFO
> Review of attachment 691139 [details] [diff] [review]:
> -----------------------------------------------------------------
> Lgtm. Not sure what you mean by 'limit that list'. Are you talking about the
> manifest?

Limit the branches list.
This is where it will run currently:
B2G xpcshell is burning on all trees apart from cedar.
Hidden on:

Once they've built it once (TBPL limitation means I cannot do so before), will need to also hide on:
I've filed bug 821249 for unhiding these once various cedar patches landed elsewhere & the runs are green.
Whiteboard: [status-b2g18:fixed]
Whiteboard: [status-b2g18:fixed]
You need to log in before you can comment on or make changes to this bug.