Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM.

Should we bump the magic number for this?

Yes, we should bump the magic number, as otherwise cached pyc files will still exhibit the bug - we need the interpreter to clear them out and recompile them.

I'm not sure this is needed. The version with the bug was not released. The bug affects only pre-alpha testers. AFAIK there are no precedences of bumping the magic number at pre-alpha stage for fixing a bug.

Fair point, especially as the new test case is the only example of the problem in the code base.

We can probably skip the NEWS entry for the same reason (i.e. it fixes a bug in never-released code, so the fact the bug existed isn't exactly relevant to most users)