Sync ruby-bunny 2.14.4-3 (universe) from Debian unstable (main)
Bug #1873058 reported by
Lucas Kanashiro
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
ruby-bunny (Ubuntu) |
Fix Released
|
Wishlist
|
Unassigned |
Bug Description
Please sync ruby-bunny 2.14.4-3 (universe) from Debian unstable (main)
This revision will fix a ruby-bunny regression in s390x. After a discussion
with one of the maintainers the way to go now is skipping this specific failing
test.
Changelog entries since current focal version 2.14.4-2:
ruby-bunny (2.14.4-3) unstable; urgency=medium
* Team upload
* d/ruby-tests.rake: skip a test consistently failing in s390x
* d/control: use secure url in Homepage field
-- Lucas Kanashiro <email address hidden> Wed, 15 Apr 2020 10:04:15 -0300
Changed in ruby-bunny (Ubuntu): | |
importance: | Undecided → Wishlist |
Changed in ruby-bunny (Ubuntu): | |
status: | New → Fix Released |
To post a comment you must log in.
The change is skipping this test on all architectures, not just s390x. And it looks like a real failure, at first glance, instead of something flaky or a race:
1) Bunny:: Queue#subscribe provides delivery handler access to message properties metadata. headers[ "participants" ]).to eq 11
Failure/Error: expect(
expected: 11
got: [11]
(compared using ==) higher_ level_api/ integration/ message_ properties_ access_ spec.rb: 72:in `block (2 levels) in <top (required)>'
# ./spec/
Once a test is being skipped, it's hard to be brought back in... Can you please file an upstream bug, link it to a launchpad one, so that we have some hope to put this test back in?
Too bad it's in debian already, there should be a link to this to-be-filed bug next to the blacklist variable, to track why a test is being skipped. Maybe you can commit something like that to debian?