finally clauses aren't treated properly in branch coverage?

[nedbat]

Brodie Rao measured cram: http://bitheap.org/cram/htmlcov/cram.html

The finally clauses are marked as partial, because they exited normally, rather than with exceptions. And the missed exit is marked as the loop exit rather than the function exit, why is that?

---

• Bitbucket: https://bitbucket.org/ned/coveragepy/issue/92
• This issue had attachments: cram.png. See the original issue for details.

[nedbat]

A simpler test case:

                for i in range(5):
                    try:
                        j = 3
                    finally:
                        f = 5
                    g = 6
                h = 7

[nedbat]

Fixed in changeset 2c5fb3a8b81c (bb).