We should not just create a result for the last occurrences of a duplicated anchor but for all of them. Else we will not be able to say to the user that anchor1 in file1 is duplicated with anchor2 in file2.

We should write a test to check if we report two results.

If we do not handle duplicated anchors well then the logic should be

• add a result all the time
• then check at the end to remove the uniquely defined ones from the results.