1

We have a team of people working on GoodData ETLs at the same time and would like to follow traditional software development practice of code review before we publish those ETLs into production. We have tried to keep those ETLs in Git-based version control system and use pull requests for code reviews but unfortunately those XML files where ETLs are stored not so human-friendly.

Are there any recommendations how to effectively review other's changes to the ETLs? Does CloudConnect provide any tools for that or maybe some special diff viewers that can better parse those ETL XMLs?

user3795752
  • 21
  • 2

1 Answers1

1

As CloudConnect is being hosted by Eclipse, it is possible to use any of the versioning system Eclipse extension, which makes collaboration on Eclipse hosted projects easier. Even though CloudConnect itself does not provide any advanced versioning systems support, the already existing Eclipse extensions can help there and the further collaboration processes can be build around them.

Please note, that some of the extensions are adding the project history sub-folder into each project’s folder, which needs to be listed in the “.deploy-ignore” file located in the root of the CloudConnect project folder in order to prevent the CloudConnect from trying to deploy this history with the rest of the process to GoodData Cloud ETL subsystem.

Filip
  • 131
  • 3
  • 1
    The problem is not really how to version control things but more about how to make reviewing changes human-friendly. Right now if one person changes ETL and another one tries to get an idea of what exactly changed it's quite hard – you just see some changes in the XML file that are double-encoded and don't immediately make much sense. So I'm trying to find better way how to essentially diff ETL changes and review them efficiently by humans. – user3795752 Jan 14 '16 at 23:50
  • I agree, also schema is a binary file, which makes it completely terrible to version control in git. – Bulat Feb 05 '16 at 09:11