Repository navigation
Tests with hexdigest assume little-endianness #425
Description
Minimal, reproducible code sample, a copy-pastable example if possible
$ pytestProblem description
When running tests on s390x (a big-endian machine), many of them fail checking the hexdigest. This is generally because the data is in a different byte order.
I don't know if tests should be creating little-endian data, or hexdigest should be changed to produce consistent results regardless of byte order.
Version and installation information
Please provide the following:
- Value of
zarr.__version__: 2.3.1 - Value of
numcodecs.__version__: 0.6.3 - Version of Python interpreter: 3.7.3
- Operating system (Linux/Windows/Mac): Fedora Rawhide
- How Zarr was installed (e.g., "using pip into virtual environment", or "using conda"): it's not; I'm building a package.
Activity
QuLogic commentedon Apr 2, 2019
Looking at least at the
test_hexdigestones, they mostly don't contain any data and are just producing a digest of the metadata. In that case, adding specific byte order to the intitialization should be fine and doesn't really affect anything. Also, when they do contain data, it gets converted to the byte order in the metadata, so that works too.Add explicit byteorder on hexdigest tests.