Jukka Lehtosalo
Jukka Lehtosalo
This is less controversial now, since in many use cases it's possible to use `Final` attributes, for which we can allow safe covariant overriding. This still wouldn't help when the...
Can we add a new error code for this check? This way we wouldn't need an additional flag.
In this case we'd need an iterator over a tuple of items. For example, `TupleIterator[int, str]` could be an iterator that can be iterated exactly twice, and the first item...
@elazarg An interesting idea! It's not strictly a tuple type since tuples support indexing with integers but this doesn't work with an iterator. However, we could perhaps add an extra...
This is now supported for functions but not for classes.
cc @msullivan @ilevkivskyi
More generally, we could analyze various open source projects and estimate frequencies of various dunder overrides in classes that inherit built-in types to prioritize which of these we might want...
This will be easier once we have a lowering pass (#854).
We can initially rely on inlining C primitive functions used for unboxing, or at least the hot paths.
I don't think the bottleneck is in `list.append`. I think that the benchmark is spending most of the time in the Python set operations. Mypyc uses a regular Python set...