Merge ~enr0n/britney:force-badtest-snapd into ~ubuntu-release/britney/+git/hints-ubuntu:impish

[Nick Rosbrook]

[William Wilson (jawn-smith)]

LGTM

[Steve Langasek (vorlon)]

The normal preferred approach now is to do a migration-reference/0 retest (--no-proposed, in retry-autopkgtest-regressions) to demonstrate that the regression in question is reproducible in the release (or -updates) pocket. Is there a reason not to do this here?

review: Needs Information

[Steve Langasek (vorlon)]

Using tools from lp:ubuntu-archive-tools:

$ retry-autopkgtest-regressions --series impish --blocks systemd --no-proposed | xargs -rn1 -P10 wget --load-cookies ~/.cache/autopkgtest.cookie -O-

I've run this now and will wait for results.

[William Wilson (jawn-smith)]

@vorlon: The reason I didn't have enr0n do that is because we know that this test is broken and we know exactly why. We'll need to SRU snapd, but are trying to get systemd to migrate in the meantime.

[Steve Langasek (vorlon)]

Ok, I'm not sure why that means you want a static hint in a file which must be edited by a member of the SRU team, vs accurate and self-updating information in the autopkgtest database that any Ubuntu developer can fix.

Historically, my threshold for adding badtest hints is "show me that the baseline has regressed, or explain to me why we should add a hint that accepts a new regression". With baseline retesting, proving that the baseline has regressed immediately means you don't have to talk to me about it.

[William Wilson (jawn-smith)]

It seems I misunderstood the functionality of the migration-reference/0 test. I thought it would just demonstrate that the tests for the package had regressed, and that a hint was still needed to not block on the regression. Looking at the excuses page though it looks like I was incorrect about that.

[Steve Langasek (vorlon)]

Cool, closing this mp then!

Unmerged commits

bd700f7... by Nick Rosbrook on 2022-04-05

ubuntu-sru: force-badtest snapd/2.54.3+21.10.1ubuntu0.2

A snapd issue is blocking many tests in impish.

diff --git a/ubuntu-sru b/ubuntu-sru
index adea1fb..123ca4b 100644
--- a/ubuntu-sru
+++ b/ubuntu-sru
@@ -3,3 +3,6 @@ force-reset-test adsys/0.7.1/arm64
 
 # This test hasn't passed in a couple of years
 force-badtest assimp/5.0.1~ds0-2/armhf
+
+# snapd regression is blocking many packages in impish
+force-badtest snapd/2.54.3+21.10.1ubuntu0.2