On Thu, 23 Sept 2021 at 19:59, Bryce Harrington
<bryce.harring...@canonical.com> wrote:
> Looking at the commit Graham points to at the upstream bug, there
> appears to be a change from using np.uint32() to .astype("<u4"). So,
> presumably issue is differences in platform datatype assumptions between
> these two routines?
In NumPy, the '<' specifies that the byte order[1] of the data type is
little-endian. So when upstream refactored that code, it ended up no
longer working on big-endian systems.
[1] https://numpy.org/doc/stable/reference/generated/numpy.dtype.byteorder.html
--
ubuntu-devel mailing list
ubuntu-devel@lists.ubuntu.com
Modify settings or unsubscribe at:
https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel
