BUG: add missing error handling in public_dtype_api.c by ngoldbaum · Pull Request #27098 · numpy/numpy

Conversation

@ngoldbaum

@SwayamInSync ran into some confusing behaviors working on a DType and I tracked it down to missing error handling here

Hard to write a test for this without defining an invalid DType.

mhvk

mhvk approved these changes Aug 3, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Makes sense. Would also be good to update the comment for the actual routine, since currently it claims to do no error checking, which is perhaps what led to this mistake:

/*
* Internal version of PyArrayInitDTypeMeta_FromSpec.
*
* See the documentation of that function for more details. Does not do any
* error checking.
* Setting priv to a nonzero value indicates that a dtypemeta is being
* initialized from inside NumPy, otherwise this function is being called by
* the public implementation.
*/
NPY_NO_EXPORT int
dtypemeta_initialize_struct_from_spec(
PyArray_DTypeMeta *DType, PyArrayDTypeMeta_Spec *spec, int priv)
{

@mattip

Labels