Gordon Wrigley
Gordon Wrigley
@ysliaw01 we did the same and have been running it in production for over a year, we do it in a base class that all our perm classes inherit off,...
@ysliaw01 I don't think it will cover it. In a two value system with object specific things like "IsOwner"(of the object) the has_permission check fundamentally has to be broader than...
I doesn't handle rst tables at all well, it'd be nice to disable it just for the rst files I have to deal with.
I'm also getting this, I'm on py3.10 with: dataclasses-json==0.5.7 marshmallow==3.16.0 marshmallow-enum==1.5.1 I observe the same behaviour around `from __future__ import annotations` Also from moving stuff around it looks like it's...
We run the pyupgrade linter which automatically updates the type hints and causes serious issues with this
Actually with light of the new day the problem we see is somewhat more aggressive. ``` ..\..\..\.venv\eligible2\lib\site-packages\marshmallow\schema.py:751: in loads return self.load(data, many=many, partial=partial, unknown=unknown) ..\..\..\.venv\eligible2\lib\site-packages\marshmallow\schema.py:717: in load return self._do_load( ..\..\..\.venv\eligible2\lib\site-packages\marshmallow\schema.py:888:...
Actually my issues are related to pyupgrade applying 3.9/3.10 features to 3.8 which can be suppressed with it's `--keep-runtime-typing` flag.
There's similar removing case to bool problems with some of the other rules, see my comment on https://github.com/MartinThoma/flake8-simplify/issues/3
I think it can only be turned into a dict literal when every element in the list is either a two-tuple or a two-list
although lists and tuples of other lengths would explode in the dict constructor, so when every element is either a tuple or list should be sufficient