Invalid classes/META-INF/fabric8/kubernetes yml files generated.

Description

We use both fabric8:helm for the charts generation with the parameters, defined in the template.yml, as well as the plain old fabric8:resource for non-helmcharts kubernetes yml I.e. template.yml: `kind: Template parameters:

• name: limits.memory value: “1000Mi”
• name: requests.memory value: “1000Mi”
• name: requests.cpu value: "200m"deployment.yml: resources: limits: memory: ${limits.memory} requests: cpu: ${requests.cpu} memory: ${requests.memory}This would become a helm template: resources: limits: memory: {{ .Values.limits.memory | default “1000Mi” }} requests: cpu: {{ .Values.requests.cpu | default “200m” }} memory: {{ .Values.requests.memory | default “1000Mi” }}` This functionality was broken in the FMP version 4.3.0, so recently we updated the plugin version to the latest 4.3.1. However this introduced a new problem. Previously target/classes/META-INF/fabric8/kubernetes/*-deployment.yml would be generated as just normal kubernetes yml files with ${limits.memory} expanded as “1000Mi” in the example above.

Now, with 4.3.1 these yml’s are generated with ${limits.memory} still in place - making them invalid.

Info

• f-m-p version : 4.3.1
• Maven version (mvn -v) : 3.6.3

• Kubernetes / OpenShift setup and version : Kubernetes 1.14.0
• If it’s a bug, how to reproduce : Per the above code snippets

[rohanKanojia]

@fforloff: poilte ping, were you able to find time to try your bug out on Eclipse JKube?

[djotanov]

I just checked with 4.3.1, those placeholders are replaced for me in classes/META-INF/fabric8/kubernetes.yml (and thats the file used in fabric8:deploy for example). @fforloff what is your use case exactly?