Jesse Glick

Please at least disable the jobs on ci.blueocean.io. As of https://github.com/jenkinsci/blueocean-plugin/pull/2379 they are less than useless—[fail in trunk](https://ci.blueocean.io/job/blueocean/job/master/) (though they were long unstable it seems), leaving a confusing red **X**...

> the digests of the files are different implying they have been fully rebuilt They have been, on GHA. This is a deliberate design choice in JEP-229 (ci.jenkins.io is considered...

> squash merging a PR with a single commit This will produce a distinct commit AFAIK, so the issue is probably limited to use of the merge queue. > also...

Oh I see, this would be coming from a `master` branch project build. I think incremental publishing can be disabled on non-PR builds; it is occasionally useful to get an...

Should be addressed going forward; not sure if you need this reopened to clean up historical duplicates between `incrementals` and `releases` @jtnord?

>> this would be coming from a `master` branch project build > > Why are these not also `-rc`? For a JEP-229 repo, the `changelist.format` is overridden to not use...

Conventionally these should be placed in a package root which is a subdirectory of `target/generated-sources` (or `target/generated-test-sources` in the case of `test` scope); for example `target/generated-sources/stuff-from-my-maven-plugin/com/mycorp/SomeClass.java`.

I would really recommend use of `target/generated-sources//` instead. This follows Maven conventions, works with `mvn clean`, and will be honored automatically by the IDE.

I am encountering a similar-sounding error while trying to update Kind (from a rather old version) running inside a GKE pod with DinD. The kernel in this case is 6.1.85...

Well, you can use the new ```groovy retry(count: 2, conditions: [agent()]) { node(…) { // … } } ``` idiom, which will retry the `node` block if it gets killed...