Damien George
Damien George
See #15997 for a fix for C6.
Thanks for attempting to fix this. The reason this hasn't been fixed before is because (1) it's difficult to do properly, and (2) it will lead to a decrease in...
I'm surprised fuzzing didn't find more cases. For example the following easily crashes the unix port: ```py @micropython.native def f(): print(x) x = 1 f() ```
I really don't think this is the right approach. There are so many combinations / cases of unbound variables being passed in as arguments. Basically every C function needs to...
> is it so that micropython.native emits bytecode and hence essentially all functions which can get called in vm.c would need a check? Not exactly. The point is, if a...
Just to follow up with some more examples that don't work with the PR in it's current state: ```py @micropython.native def f(): x = 0 print(x, y) y = 1...
Righto. Well, if it's supposed to be that way, then it's going to be a hack to implement.
Seems that `__add__` is correct (ie goes to built in). Also seems that explicit `t.__str__()` calls Foo2's (so "incorrect").
And yet `__repr__` will go to the tuple! (via both `repr(t)` and `t.__repr__`).
I get: ``` >>> list.__str__ >>> list.__repr__ ```