flake8-abp doesn't recognize "from __future__ import unicode_literals"

[sebastian]

Background

​flake8 is a Python code quality checker which we recently started to use along our own extension ​flake8-abp. One of the things flake8-abp does is making sure that string literals comply to ​our coding style, by comparing them to the output of repr() in Python 2, or ascii() in Python 3. However, the current implementation doesn't consider from __future__ import unicode_literals, which if given at the top of the script, causes string literals (without b prefix) to be interpreted as unicode strings, just like in Python 3.

How to reproduce

Run flake8 (using Python 2 and with flake8-abp installed) on following script:

from __future__ import unicode_literals


foo = '\u1234'

Observed behaviour

flake8 outputs following warning:

stdin:4:7: A110 string literal doesn't match repr()

That is because since from __future__ import unicode_literals is ignored the string is interpreted as a byte string where \u has no special meaning and therefore repr('\u1234') returns '\\u1234'.

Expected behaviour

If from __future__ import unicode_literals is used in the file, flake8-abp should consider all string literals (without b) prefix unicode strings, and shouldn't throw a warning in the example above.

[jsonesen]

Do you think that it makes sense to add the logic for handling the import to the 'check_quotes' method that raises the exception?

[abpbot]

A commit referencing this issue has landed:
​Issue 4044 - Added handling for __future__ unicode_literals import to check_quotes()