How to deal with automated scripts when multiple releases are in progress simultaneously

Question

We have a continuous development architecture where developers are working on multiple software releases at the same time like developing/ fixing R6.5 or R5.8 releases simultaneously. Sometimes there are major GUI differences between the two releases. I need help on how to handle objects/object properties in such kind of multi release environments where we have major GUI differences between two different releases and we need to test both the releases simultaneously.

Answer 1

Assuming that you use a version control system like Git or Subversion for the source code of your application, my suggestion would be to use the same source control for your test scripts as well. That way, whichever branching and versioning strategy you decide to use for your application can apply also to the test scripts.

This is of course only possible if your test tool supports working with external version control systems. If you use a tool like Silk4J, Silk4NET, Silk Test Classic, LeanFT, or Selenium then this shouldn’t be a problem as those tools store the actual test scripts as plain text files in the file system, which can be versioned with whichever source control system you choose.

If you use a tool like the Silk Test Workbench, things gets a little bit more complicated as the scripts are stored in a separate data base. In that case, I would recommend trying to replicate the version structure of your application in such a test tool as well. For example, if you have a "trunk" of your application and various release branches (that you may work on in parallel) I would recommend creating a "trunk" test project in the Silk Test Workbench and to create a copy of that project for each release "branch". That way you’ll have multiple copies of your test scripts (including Object Maps, and other assets) which introduces a slight overhead in terms of maintainability, but you’ll be able to run and modify the test scrips independently of each other, depending on the peculiarities of each individual release.