The value of regression testing is often not realized until it is automated. After that, everything will become much easier.
This means that you should be able to start from a known position (if you create MD5 on files, you must start from the same files every time). Then get one successful run in which you can save the result - this is the basic level.
From now on, regression testing is just a button task. Start your test, collect the result and compare it with a known base (of course, if the output will ever change, you will need to check it manually or using another independent script before saving it as a new baseline).
Remember that the idea of โโregression testing is to catch any errors introduced by the new code (i.e., regress the software). This does not test the functionality of this new code.
The more you can automate this, the better, even as a one-person development team.
source share