Bug 1854254

Summary: Build fails with evolution-data-server 3.37+ due to test failure: EDSBackend / add-persona FAIL (killed by signal 5 SIGTRAP)
Product: [Fedora] Fedora Reporter: Adam Williamson <awilliam>
Component: folksAssignee: Brian Pepple <bdpepple>
Status: CLOSED RAWHIDE QA Contact: Fedora Extras Quality Assurance <extras-qa>
Severity: urgent Docs Contact:
Priority: unspecified    
Version: rawhideCC: bdpepple, debarshir, kevin, mcrha, philip
Target Milestone: ---   
Target Release: ---   
Hardware: All   
OS: Linux   
Whiteboard:
Fixed In Version: folks-0.14.0-2 Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2020-07-07 12:58:52 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Attachments:
Description Flags
full test log file from failed build on F32 with e-d-s 3.37.1 none

Description Adam Williamson 2020-07-07 00:30:32 UTC
Attempting to build folks in current Fedora Rawhide fails due to this test failure:

30/44 folks:EDSBackend / add-persona               FAIL           7.03s (killed by signal 5 SIGTRAP)

the logs don't really make it at all clear what that means exactly. Is it a normal test failure? Is it a test system issue?

This is a major problem because someone tried to bump evolution-data-server in Rawhide, but we had to untag that bump and a large number of rebuilds against it because folks cannot be rebuilt due to this bug, and that breaks Workstation image composes, which breaks Rawhide composes.

Comment 1 Adam Williamson 2020-07-07 01:05:03 UTC
So I'm pretty sure this is caused by e-d-s 3.37 somehow.

I tested building latest folks on Fedora 32, works. Then I rebuilt evolution-data-server-3.37.1-1.fc33.src.rpm on Fedora 32, added it to a side repo, and tried building folks again, and it fails in the same way as Rawhide:

30/44 folks:EDSBackend / add-persona               FAIL           6.79s (killed by signal 5 SIGTRAP)

so, build on F32 works, build on F32 but with e-d-s 3.37.1-1 fails.

I'll attach the full test output log from that build attempt.

Comment 2 Adam Williamson 2020-07-07 01:05:55 UTC
Created attachment 1700090 [details]
full test log file from failed build on F32 with e-d-s 3.37.1

Comment 3 Adam Williamson 2020-07-07 01:10:28 UTC
so the build has likely been broken ever since 3.37.1 landed in Rawhide, but since 3.37.1 didn't actually bump any sonames compared to 3.36, it wasn't an immediate problem. we only noticed when 3.37.3 landed, because that *does* bump sonames.

Comment 4 Milan Crha 2020-07-07 07:25:37 UTC
Weird. I tried a local build and I'm pretty sure it used my local build of eds, and the very same test passed with no problem. I'm not sure how to debug the folks test, I do not speak vala, but I'll try at least.

Comment 5 Milan Crha 2020-07-07 09:13:41 UTC
This took me a bit... the reason is the added libphonenumber dependency. The test fails to find phone-1 and phone-2 attributes, because they have added parameter x-evolution-e164, which is due to the libphonenumber usage. I'll propose an upstream patch for it.

Comment 6 Milan Crha 2020-07-07 09:31:06 UTC
A change proposal for folks is here:
https://gitlab.gnome.org/GNOME/folks/-/merge_requests/40

Comment 7 Milan Crha 2020-07-07 12:58:52 UTC
And here is the backported change:
https://src.fedoraproject.org/rpms/folks/c/d5eb016e282b16eb92d9526ea0b02e39bf294284?branch=master

It's already built in the side tag "f33-build-side-25060" and will be part of rawhide soon.

Comment 8 Adam Williamson 2020-07-07 15:13:16 UTC
yikes, fun! nice catch :) thanks for the fix.