Closed Bug 852543 Opened 10 years ago Closed 10 years ago

B2G desktop builds failing during make check on m-c


(Firefox OS Graveyard :: General, defect)

Not set


(Not tracked)



(Reporter: RyanVM, Assigned: fabrice)




(1 file)

Hiding them (again) for now...

XPCOM_DEBUG_BREAK=stack-and-abort /builds/slave/m-cen-osx64_g-0000000000000000/build/obj-firefox/dist/bin/ \
	  ../../dist/bin/TestRegistrationOrder "/builds/slave/m-cen-osx64_g-0000000000000000/build/xpcom/tests/regorder";
Running RegistrationOrder tests...
TEST-PASS | TestRegular
TEST-UNEXPECTED-FAIL | TestJar FAILED - cannot create extension service

TEST-PASS | RegTestOrderC
Finished running RegistrationOrder tests.
Adding some XPCOM peers to the party. This needs an owner, it's blocking B2G desktop builds from running on m-c and inbound.
Severity: normal → critical
Is extension2.jar present and has the same/correct contents as in other builds?
How would I tell? As best I can tell, those tests don't get packaged?
extension2.jar is part of the source code ( I don't know how the B2G config differs from a regular config... perhaps the source is not at the normal expected location?
These are builds off m-c and inbound, so extension2.jar is in the same place.
FWIW, we don't run make check on the b2g18 branches, so it's possible that this is broken there too and we just haven't known about it.
Yeah, I didn't mean that the source location was different, but perhaps the location of the entire B2G source tree is different?
AFAICT, these builds use the regular m-c source tree with Gaia cloned into $topsrcdir/gaia

Andreas, can you suggest someone who could help debug this?
Flags: needinfo?(gal)
So, checktests were disabled on purpose for the b2g18 branches in bug 821401 comment 20. Unfortunately, nobody ever followed up on fixing said tests so that we could enable them again. So here we are. Are we going to disable our way to victory again or can we look into what it takes to get this fixed and these tests enabled?
Flags: needinfo?(jhford)
I can reproduce on a linux build also, I'll dig into that.
Assignee: nobody → fabrice
Attached patch patchSplinter Review
Kyle, the patch is in the binary jar, but is just applying there the same change as the one at:
Attachment #727906 - Flags: review?(khuey)
Attachment #727906 - Flags: review?(khuey) → review+
Closed: 10 years ago
Resolution: --- → FIXED
Flags: needinfo?(jhford)
Flags: needinfo?(gal)
Blocks: 856248
You need to log in before you can comment on or make changes to this bug.