I do exactly the same. Added to this, in large legacy code bases with no or very little tests, i sometimes try and break the functionality near to the part wher

I do exactly the same. Added to this, in large legacy code bases with no or very little tests, i sometimes try and break the functionality near to the part where a new fix is required in my private branch to check my understanding. I find that its much easier to understand the code when something doesn't work rather than when its working perfectly. With any luck, the stacktrace will tell me what's going on.