We've encountered a problem where the shape of one of the dimensions of a numpy array was set to be >2^32, which it shouldn't be. This was solved by casting to 32-bit, which would then ignore the 32 msb bits. However, this cannot be done for all of the struct values from Python, as some of the sizes are larger than 32-bit.
To solve this, we should implement macros for proper handling of these values. It would also have the added benefit of less boilerplate code.
We've encountered a problem where the shape of one of the dimensions of a numpy array was set to be >2^32, which it shouldn't be. This was solved by casting to 32-bit, which would then ignore the 32 msb bits. However, this cannot be done for all of the struct values from Python, as some of the sizes are larger than 32-bit.
To solve this, we should implement macros for proper handling of these values. It would also have the added benefit of less boilerplate code.