Stack overflow caused by clumsy array coercion #17785
Description
Reproducing code example:
Over at pygae/clifford#376, we've found that the adjustments to array coercion in 1.19.x seem to produce stack overflows where previously they were safe. A minimal repro is this clumsy coercion code, which mirrors what clifford.MVArray did in their 1.3.0 release.
import numpy as np
class MyArray(np.ndarray):
def __new__(cls, input_array):
obj = np.empty(len(input_array), dtype=object)
obj[:] = input_array
obj = obj.view(cls)
return obj
class MyNastyClass:
def __len__(self): return 32
def __getitem__(self, item): return 1
def __array__(self): return MyArray([self])
np.asarray(MyNastyClass())Error message:
Windows fatal exception: stack overflow
Current thread 0x00006160 (most recent call first):
File "<ipython-input-4-ac7cb36f8e78>", line 3 in __new__
File "<ipython-input-4-ac7cb36f8e78>", line 12 in __array__
File "<ipython-input-4-ac7cb36f8e78>", line 4 in __new__
File "<ipython-input-4-ac7cb36f8e78>", line 12 in __array__
File "<ipython-input-4-ac7cb36f8e78>", line 4 in __new__
File "<ipython-input-4-ac7cb36f8e78>", line 12 in __array__
NumPy/Python version information:
1.19.4 3.8.0 (tags/v3.8.0:fa919fd, Oct 14 2019, 19:37:50) [MSC v.1916 64 bit (AMD64)]
Thankfully the code above was removed since it was bad for other reasons, but its unfortunate that users on an old release get a segfault.