-
-
Notifications
You must be signed in to change notification settings - Fork 85
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Test Failures on s390x #467
Comments
Hi. I think we have never tried C-Blosc2 with s390x arch before. Also, a cursory look into your output does not point to something immediately apparent to me. In case you find the issue, we would be happy to incorporate the fix. Thanks. |
Hi @FrancescAlted Do you know if c-blosc2 has the capability to work on big-endian architectures? If so, could you give some pointers to where this is handled for other (supported) big-endian archs? I would be happy to submit a PR if we can get this working. |
Hi @ELundby45 . Well, C-Blosc2 should be big-endian friendly, except when reading a frame from disk. This part starts at the comment line '// TODO: make this portable across different endianness` and was introduced in commit: 8b5ea6f#diff-5de69465579e00aa53d6ff4e7fb060d8c08a03aa1721255fce9de3afbb27ba1a The issue appears to be that the values in the chunk index are written in little-endian, but there is no protection for big-endian machines when reading them back. For more information on the chunk index, see https://github.com/Blosc/c-blosc2/blob/main/README_CFRAME_FORMAT.rst#chunks and search for |
I see something similar. The tests are very unhappy on s390x:
|
If needed I can provide access to a s390x system for development/debugging. |
Same issue on debian/sid |
+1 for avalentino also fails with a newer version 2.13.1: as well as on Ubuntu with 2.13.1+ds-1ubuntu1: |
A friendly reminder that this is a known issue, and help would be appreciated to fix it. For details, see #467 (comment). |
I am trying to build a Conda package for c-blosc2. It is passing on all architectures except for IBM-Z (linux s390x), which is failing due to tests not passing. I have 2 questions:
Steps to reproduce:
Run conda build on s390x
Expected behavior:
Tests pass
Full build log:
c-blosc2 s390x failure.log
GCC compiler
linux-s390x
Version 2.8.0
The text was updated successfully, but these errors were encountered: