Ewen Cheslack-Postava commented on KAFKA-5939:

I think dry run is fine if we're clear about what it means. The scariest part 
of developing the script originally was the final steps around pushing tags. 
Otherwise most of it is only affecting stuff that's easy to clean up. To me, 
the most useful dry run would:
 * Still prompt about the steps that would upload, but just say what they would 
 * Say what tagging it would do, but skip it
 * Still do the full build and provide artifacts, which allows you to do some 
test run and validation without really generating an RC

I'm not sure how the publication to the maven repo would work as a dry-run, 
maybe just publish to a local directory.

The other thing which could possibly be handled here or could be a different 
Jira is to have a local copy of all artifacts when you've completed. This would 
apply to the regular release artifacts (it was a pain to pull them back down 
from my Kafka home directory in order to promote the release) and the maven 
artifacts (which is useful for doing validation, which should also be 

> Add a dryrun option to release.py
> ---------------------------------
>                 Key: KAFKA-5939
>                 URL: https://issues.apache.org/jira/browse/KAFKA-5939
>             Project: Kafka
>          Issue Type: New Feature
>          Components: tools
>            Reporter: Damian Guy
>            Priority: Major
> It would be great to add a `dryrun` feature to `release.py` so that it can be 
> used to test changes to the scripts etc. At the moment you need to make sure 
> all JIRAs are closed for the release, have no uncommited changes etc, which 
> is a bit of a hassle when you just want to test a change you've made to the 
> script. There may be other things that need to be skipped, too

This message was sent by Atlassian JIRA

Reply via email to