Wilfredo Sánchez Vega
Wilfredo Sánchez Vega
This is solvable with a custom `__init__` as per #393: ```python @attr.s(extra_args=True) class Child(NonAttrsClass): a = attr.ib() b = attr.ib() def __init__(self, a, b, c): self. __attrs_init__(self, a, b) #...
Yeah… I'm just wonder out loud about how many variations of how one could change `__init__`'s handling of args we want to build in versus just allowing one to take...
Back to @pbourguignon's proposal, would it be feasible to just notice extra arguments in the signature for `__attrs_post_init__`? As in: ```python @attr.s() class Child(NonAttrsClass): a: int b: str def __attrs_post_init__(self,...
> On the other hand, the solution you suggest has the added benefit of `__init__` signature being complete. I think that's pretty important for typing.
@hynek so… out of scope for `attrs`…?
The `t.i = None` case can be caught now with the `on_setattr` hook.
Any action left here?
I think raising an `ValueError` or similar is more appropriate than sanitizing the given data. Sanitizing the data is likely to just delay an error to a less-obviously-related-to-the-problem part of...
Right, but you phrased the initial description as a JSON problem. :-) Mapping JSON schema names to python object attributes is what's not `attrs`' problem. But I think I understand...
Sorry, I thought I'd have this tons-of-free-time summer this year and yet reality disliked my plans. Anyway, I haven't been following along. Is there an example of that this is...