Re: [che-dev] Che stack for Apache Camel/Fuse projects

[Yevhen Ivantsov <yivantso@xxxxxxxxxx>]

Hi Aurelien!

What are the requirements to have it directly integrated in Eclipse Che?

There are no requirements as such. We used to accept PRs into stacks.json and samples.json, whoever ended up in having a lot of stacks and samples that do not work. Of course this isn't good UX. Thus, we left only basic stacks.

How tests are handled for Che stacks? (there are no tests in the community project)

There are stack tests here https://github.com/eclipse/che/tree/master/selenium/che-selenium-test/src/test/java/org/eclipse/che/selenium/stack

What are the requirements to have the Camel Docker file in this repository?

Dockerfiles in that repo are used for automated builds on DockerHub to produce eclipse/$image images. To be honest, there are no reqs, now we try to keep the default stacks and sample list as soon as possible.

Camel requires the same kind of dependency with SpringBoot but with a different version. What is the best:

SpringBoot image is just a Java image with ~/.m2 pre-populated with Springboot deps.  It is usually done to match Springboot version of a related sample application.  So, if you update deps there, you should probably update version of Springboot the sample app uses.

I also noticed that we can have single-machine or multi-machine. Several questions comes to my mind:

There are no words of caution regarding multi machine workspaces. The only thing to bear in mind is that recipe type: compose isn't supported on k8s/OpenShift infra and vice versa. Memory consumption should be the same.