Conversation
|
Welcome @mk24x7! |
|
Hi @mk24x7. Thanks for your PR. I'm waiting for a kubernetes member to verify that this patch is reasonable to test. If it is, they should reply with Regular contributors should join the org to skip this step. Once the patch is verified, the new status will be reflected by the I understand the commands that are listed here. DetailsInstructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository. |
|
Can one of the admins verify this patch? |
edba5bb to
e5f3105
Compare
nirs
left a comment
There was a problem hiding this comment.
I will help if you post first the workflows changes so I can see the linter errors. Or share the errors from CI in a comment.
|
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: mk24x7 The full list of commands accepted by this bot can be found here. DetailsNeeds approval from an approver in each of these files:Approvers can indicate their approval by writing |
|
@mk24x7 The windows lint does not work. You can test locally or in your fork. It will be good enough to remove the windows lint from this PR since it requires additional work. We can open a new issue to make Another solution is to use the golangci-lint action intead of running it manually. I'm not sure make is the right solution for windows. @bobsira can you help with this? What it the best way to handle this on Windows? |
This seems like an easy fix, use the standard name pattern?
Yes, the kvm driver is not compiled on arm for years. We need to use a special cross compiler configuration to make it work and we did not have time to work on it yet. For now, we can fix the tags to match the build and avoid lint errors.
You can start by removing it from the matrix so we can merge the easy fixes now and open an issue for fixing kvm lint issues. |
|
@mk24x7 Also update the PR description to describe the current change. |
|
@mk24x7 The windows unit tests fail - maybe windows lint issue are wrong? Please remove them from this PR since we don't have a lint job to verify them yet. |
fd04696 to
8418f99
Compare
The Lint job ran only on ubuntu-latest, so build tags hid lint issues in code guarded by //go:build darwin or windows. Those issues surfaced only on developer machines. Run the job on a matrix of os/arch variants instead: ubuntu-24.04, macos-26 and macos-26-intel. fail-fast is disabled so one failing variant does not mask the others, and the lint step runs under bash so the same command works everywhere. windows-2025 and ubuntu-24.04-arm are skipped for now. The kvm driver is not built on arm64, so linting it there reports unused code that is used on amd64. make lint does not work on Windows yet, because the golangci-lint download assumes a tar.gz archive while the release publishes a zip for Windows. Both will be handled in follow up issues. Adding the jobs first surfaces the issues that need fixing. Fixes for those follow in later commits.
These are in code guarded by darwin build tags, so the linux-only lint job never saw them. vfkit: the first two writes have no formatting arguments, so use fmt.Fprint instead of fmt.Fprintf. krunkit: the unchecked type assertion on the error returned by vmnet.ValidateHelper panics if the error is not a *vmnet.Error. Use errors.As and fall back to the original error when it does not match. parallels: the unchecked type assertion on the driver panics if the type ever changes. Check it and return an error instead, matching the existing pattern in the vmware driver.
8418f99 to
8f2ed8b
Compare
8f2ed8b to
6064556
Compare
|
Removed the windows lint fixes and windows-2025 from this PR, and updated the description. It is now the matrix on ubuntu and macos, plus the three macos fixes. How would you like to sequence this with #23776? Happy to hold this until that is fixed, or to drop the workflow change here and leave just the macos fixes, whichever works better for you. |
|
|
@mk24x7 Note the failure on macos-26-intel: The inel runners are much slower so we need bigger timeouts. This failure may not be relevant after #23782. |
Run the Lint job on more os/arch variants, so lint issues in code behind build tags are caught in CI.
The Lint job ran only on ubuntu-latest, so issues in code guarded by
//go:build darwinwere never linted. The job now runs on ubuntu-24.04, macos-26 and macos-26-intel, withfail-fastdisabled so one variant does not mask the others.The macOS commit fixes what this revealed:
pkg/drivers/vfkit/vfkit.go:fmt.Fprintfwith no format argumentspkg/minikube/registry/drvs/krunkit/krunkit.go: unchecked type assertion that could panicpkg/minikube/registry/drvs/parallels/parallels.go: unchecked type assertion that could panic, using the same pattern as the vmware driverNot included, per review:
Fixes #23746