Hi Karen! Excellent question. Generally, to test changes to the precommit patch tester I've seen folks make patches (using git format-patch) that have two commits, one for their change and one that alters something related to what they're changing.
When such a patch shows up attached to a JIRA, precommit happily applies both changes and then runs against them (incorporating most changes to test patch itself in the process). If that won't work, then you can safely run test-patch on a local machine against an arbitrary project checkout. With default options it shouldn't reset the local source control working directory. So long as you don't give it login credentials it won't post to JIRA, GitHub, etc. If you're trying to reproduce a particular project's use of precommit, then one of the existing committers with access to builds.apache.org can get you the invocation details said project uses. Do either of those options help in this case? On Aug 10, 2016 20:22, "Karen Clark" <[email protected]> wrote: > Howdy folks, > > I'm working on a trivial bug (YETUS-342) to update test-patch.sh to > gracefully handle the situation where $BASEDIR does not exist. > > Now that I have a fix, I need to ensure I didn't break anything (and to > ensure the fix works in various situations). > > Unfortunately, I'm not developing anything in the Hadoop ecosystem (or > elsewhere!) that I can run my changes on to validate them using Yetus. > > Is there sample data available that we can test the code against without > fear of committing anything? > > Thanks, > Karen >
