New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Pull in latest snapd trunk and add additional optional parameter as nil to gadgetLayoutVolume #136
Conversation
…il to gadgetLayoutVolume.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
nice
Ok, the test failures now seem to be pointing at a real issue - possibly testcase needs to be readjusted. I think something changed in snapd. |
I'm a bit worried about the gob tests that are failing right now. Apparently the latest changes cause issues with state machine resume... |
The gob tests fail because type kcmdline.ArgumentPattern has no exported fields. This is introduced in snapcore/snapd#12976 |
Are there some more details on the "gob" test failure? We have a fix in snapcore/snapd#13144 but it would be nice to understand what exactly fails on your side so that we can add a test in snapd to prevent this in the future. |
The CI shows
The error happens in readMetadata, ubuntu-image/internal/statemachine/state_machine.go Lines 333 to 348 in 28b7b85
Actually the error should happen early, when encode the data. But the error is ignored, so I send #138, then it will tells that writeMetadata also fails. ubuntu-image/internal/statemachine/state_machine.go Lines 369 to 380 in 28b7b85
ubuntu-image use ubuntu-image/internal/statemachine/state_machine.go Lines 95 to 125 in 28b7b85
And it wants to use encoding/gob to serialize that struct. So if snapd wants to add a test, then it could try to use encoding/gob on the |
Is there a way we could make the code fail to compile if it is not serializable rather than getting a nil? It feels like we would need to write test with lots of different input so that we test all possible types. And that is not nice. |
Maybe we should implement GobEncoder and GobDecoder on |
hm, sadly I'm seeing issues in our CI with the latest snapd version.
It didn't happen with the previous version, and it's not even in the pause/resume tests but instead in the UC20 tests. I can reproduce it locally as well. I'm trying to investigate, but would appreciate some help. |
Merging this, as it seems to make everything and everyone happy. |
This should fix snapd integration test suite, which got broken after snapd trunk changed the API of
gadget.LayoutVolume
. It's a bit of a shame that because of such API changes we need to rebuild with snapd trunk.