[ https://issues.jboss.org/browse/JBIDE-24682?page=com.atlassian.jira.plugi... ] Hardy Ferentschik commented on JBIDE-24682: ------------------------------------------- {quote} Escalating this issue. Ondrej Dockal Really good catch. Thanks for being so on-the-ball here. {quote} +1. This is really a regression on side of the Minishift/CDK team. We need to see what the best way forward is on this one. For what it's worth, the change comes from this issue - https://github.com/minishift/minishift/issues/924 - which originated from the community. The version output of Minishift differed cute substantially from the version output of related tools like Minikube, oc, OpenShift, etc. We aligned the Minishift version string, forgetting that it is used by JBIDE server adapter. We had no unit test (and/or comment) which warned us from making a change without reaching out first. The unit test exists now (against the new format) and we will for sure add a comment to the code moving forward. {quote} Output of 'minishift version' was (and is) expected to follow a consistent format at the bare minimum. {quote} Correct. {quote} If we accept in CDK 3 server adapter both cdk and minishift binary maybe we should create a CDK jira that would ask for changing output from 'minishift version' from 'CDK Version: X.X.X' to 'cdk vX.X.X'. I do not know if we plan to do so, it is just an idea... {quote} That would be my suggestion as well. Right now the format is not consistent, but it should be. It seems people prefer the format without colon, so I suggest to make the CDK version string consistent with the current Minshift format. For the server adapter it would probably mean to split the lines at the whitespace, in order to get the key value pairs. -- This message was sent by Atlassian JIRA (v7.2.3#72005)