| 4 | |
| 5 | The patch provided by the reporter is closer to the correct solution. However there are many more strings is this module that became unicode with `unicode_literals`, with unknown consequences. |
| 6 | |
| 7 | We have several options available now, by order of increasing effort: |
| 8 | - roll back `unicode_literals`, and fix whatever needs fixing for Python 3, |
| 9 | - liberally sprinkle `str('..')` in the module, and hope we didn't miss one, |
| 10 | - as suggested in #19081, compare the current code with the 2.6 and 2.7 stdlib, and switch to the stdlib wherever possible, subclassing if necessary. This should make the issue go away. |