During our daily work to improve our Gupta migration C# library (the PPJ Framework) and support to partners working on projects, we usually encounter a good number of bugs trying to migrate to .NET. They are coming from the customers' original Gupta code and have been hidden for a long time.
In many cases those bugs have been hiding for years behind "features" of Team Developer that someone may consider bugs. For example: The edited flag (retrieved using SalQueryFieldEdit) is not cleared when setting the value of a cell that doesn't have the focus. It has been like this in Gupta forever. Bug or feature?
But the real question for us is: Do we replicate this "feature" in our implementation because the original code may rely on the fact that the modified flag in cells is almost never cleared?
Some developers argued that we should replicate everything, including strange behavior like this one.
We have decided not to migrate bugs. We decided to make the PPJ Framework a highly reliable .NET implementation of most of the features found in Team Developer without all the known and unknown bugs (when we discover them).
As a result, we have to be able to detect and change the original code that may be affected by the fix. It's a little more work, but the end product is a more robust and reliable .NET application.