I'm working on automating a multi-project build with the NetKernel Gradle plugin. I approached the job in two stages:

1. Create gradle build definition and test build/deploy tasks in/out in my local (Windows machine)

2. Thaw and run previously frozen NetKernel instance in a deployment environment (Linux machine)

My problem derives from the fact that my local is a Windows machine while the deployment environment is a gcloud instance that runs Ubuntu.

Running following command in the *nix instance: > ./gradlew :startTEST -PtestInstanceLocation=/opt/netkernel/NKSE-5.2.1-test

reports following error:

Cannot run program "/opt/netkernel/NKSE-5.2.1-test/bin/netkernel.sh" (in directory "/opt/netkernel/NKSE-5.2.1-test"): error=2, No such file or directory

... for a perfectly valid reason: netkernel.sh does not exist indeed as the frozen NK instance contains netkernel.bat (frozen instance being initially installed on Windows).

Given the situation above, I suggest the NetKernel installer should generate both startup script variants for Windows and *nix (.bat and .sh).

Note: I incline to say NetKernel installer should generate both startup script variants when run on Windows. But what if someone has *nix in a source machine (local) and Windows on target (deployment machine). Quite rare, indeed...

Do you find my suggestion reasonable?

PS In my local Windows, I run shell terminals as well. Having .sh startup scripts would help on that front as well.

PS2 If I remember correctly, Tomcat generates/packages both .bat and .sh startup scripts.

Best regards, mircea