Renovate does not suggest updates to k3a-embedded

[sverrehu]

Perhaps renovate does not understand how to handle semver 2 metadata.

[erikgb]

Would it be possible to use Kafka version as a classifier in k3a-embedded?

[sverrehu]

This is what the Renovate log says in a local run:

 {
  "datasource": "maven",
  "depName": "no.shhsoft:k3a-embedded",
  "currentValue": "0.2.1+${kafka.version}",
  "fileReplacePosition": 909,
  "registryUrls": [
    "https://repo.maven.apache.org/maven2"
  ],
  "depType": "test",
  "skipReason": "version-placeholder",
  "depIndex": 2,
  "updates": []
}

Note the "skipReason": "version-placeholder". Not sure changing to classifier will solve this.

[erikgb]

@sverrehu Could it make sense to publish a version of k3a-embedded without the build number (ref. SemVer), that is built with the latest version of Kafka available at the time the release was cut? I know it is not that "clean", but I think it should work quite well for us in this project.

[sverrehu]

@sverrehu Could it make sense to publish a version of k3a-embedded without the build number (ref. SemVer), that is built with the latest version of Kafka available at the time the release was cut? I know it is not that "clean", but I think it should work quite well for us in this project.

I used to do that, but disabled it for the last release just because of the cleanlyness you mention.

[erikgb]

@sverrehu Could it make sense to publish a version of k3a-embedded without the build number (ref. SemVer), that is built with the latest version of Kafka available at the time the release was cut? I know it is not that "clean", but I think it should work quite well for us in this project.

I used to do that, but disabled it for the last release just because of the cleanlyness you mention.

But this build number thing does not work well - that we know (e.g. this issue). A classifier was already suggested, which I think might work better from a user point of view, but.....

[sverrehu]

@sverrehu Could it make sense to publish a version of k3a-embedded without the build number (ref. SemVer), that is built with the latest version of Kafka available at the time the release was cut? I know it is not that "clean", but I think it should work quite well for us in this project.

I used to do that, but disabled it for the last release just because of the cleanlyness you mention.

But this build number thing does not work well - that we know (e.g. this issue). A classifier was already suggested, which I think might work better from a user point of view, but.....

Trying to figure out how to do a revert on GitHub. :-)

It is available on the PR, and not the commit. :-)