Repository navigation
Making it easier to use unicode_literals #2536
Description
Currently using from __future__ import unicode_literals in Python 2 is kind of painful with mypy. Here are some potential issues:
- Some typeshed stubs declare things to only accept
strarguments, when in reality they also accept at least ascii-onlyunicodeobjects. - Mypy might not accept the implicit
unicodeliterals in every context where a literal has a syntactic role (outside normal expressions), andb'foo'literals won't work in Python 3. String literal type escaping is an example.
It's fair to argue that using unicode_literals is a bad idea, but it's not uncommon to encounter real-world code using that, so getting mypy to work with it might be at least somewhat important.
Activity
ddfisher commentedon Dec 7, 2016
FWIW string literal type escaping has already been fixed: #2516
After that fix I did a quick grep of the codebase looking for other similar errors, and the only one I found is fixed in #2519.
JukkaL commentedon Dec 7, 2016
This is still rejected by mypy when using
--py2 --fast-parser:JukkaL commentedon Jan 29, 2020
This doesn't seem worth fixing any more, since Python 2 reached it's EoL and gradually the amount of code using
unicode_literalsis likely to decrease.