Closed convexquad closed 7 years ago
Don't include jline in azkaban-client build script as it overrides older version of jline included in hadoop-plugin. It caused hadoop validator to not work previously.
Whoops, I had stashed these changes from an earlier point in the commit history and forgot to rebase. I will remove jline from azkaban-client.
Oh, sorry. You have already taken care of it. I would want to test if Hadoop Validator works appropriately before pushing this version.
What do you use for testing the Hadoop Validator tasks? Can I just run the hadoopValidate task in the Hadoop Starter Kit?
Yes, you can. Follow go/hadoopvalidator to setup hadoop-starter-kit to be able to run it. I'm really excited to use this enhancement soon. I would use it once you push this. 💯
Never mind, as you removed jline dependency for azkaban-client it would anyway work.
In the hadoop-plugin project I moved jline:jline:1.0 to a runtime dependency, since it isn't actually used for compilation (but is indeed necessary at runtime).
I tested all the tasks with the hello-pig-project in the HSK and they all work correctly. The pigDataExists task fails since one of the datasets used in the HSK is no longer available (the old MEMBER2 database).
I am going to merge this PR since my SNAPSHOT build works correctly with the HSK. This PR will help me complete my work on the Hadoop DSL Intellij syntax completion.
@pranayhasan The new pigValidate tasks are totally awesome! I am super impressed with them. It is almost Thanksgiving here in the US, but after I come back I will help promote these new tasks, and also I want to do a bit more refactoring with the code base so that we can keep the tasks in good shape moving forward.
…plugins to all projects