You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The other day we got a dependabot PR bumping go-containerregistry #2664. Unfortunately the integration tests in Azure Pipelines did not kick off at all AND we didn't have our branch protection rule set to require porter-integration as a status check so it looked like the PR was safe to merge.
Narrator: It was not.
Bumping go-containerregistry has a breaking change that caused our archive files to change their hash. This is causing TestArchive_StableDigest to fail because the expected hash is different.
For now we need to revert that change, and then evaluate what changed in go-containerregistry and how that impacts our archives, and if there's a way to not have the hash of our archives change. Otherwise we'll need to communicate to people that it's changing and when, since I anticipate that some users are checking the tgz hash when moving across an airgap.
The text was updated successfully, but these errors were encountered:
@schristoff What do we want to do here? The dependency have been updated multiple times since v1.1.5, we are now using v1.1.12. The time for avoiding the change is gone, as we have released two versions of Porter with the the updated dependency.
The other day we got a dependabot PR bumping go-containerregistry #2664. Unfortunately the integration tests in Azure Pipelines did not kick off at all AND we didn't have our branch protection rule set to require
porter-integration
as a status check so it looked like the PR was safe to merge.Narrator: It was not.
Bumping go-containerregistry has a breaking change that caused our archive files to change their hash. This is causing TestArchive_StableDigest to fail because the expected hash is different.
For now we need to revert that change, and then evaluate what changed in go-containerregistry and how that impacts our archives, and if there's a way to not have the hash of our archives change. Otherwise we'll need to communicate to people that it's changing and when, since I anticipate that some users are checking the tgz hash when moving across an airgap.
The text was updated successfully, but these errors were encountered: