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
MRG, BUG: Explicit read endianness #8618
Conversation
First of all, thanks for the speedy response <3 I have a small doubt regarding the tests. The tests are first writing the test data in the specific encoding and then the same test data is being read during the tests? If this is case, then it makes sense to fix the encoding while writing the files. Other than that, the tests are currently failing, mainly because certain functions try to read the files directly in little-endian encoding. For example: egi.py, unpack in kit I also ran a build with this branch which failed which I can see from the above reasons. Below are the details about the build. The full log is here
|
In general, no, we don't write files then read them for all formats. We only really do this for FIF. I see the |
Can you try again @Aniket-Pradhan ? |
There's one more unpack call in mne/io/kit/coreg.py. The tests are passing now, I guess once the above unpack call is fixed, it should be ready to go then. |
Almost all the failing tests seem to be working well now. Some tests still fail on the s390x arch, which I guess again might be because of the endian-ness, but am not sure. Can you have a look at them? Here's the complete build log. Below's the info on the failing tests:
|
Okay caught one more, can you try again @Aniket-Pradhan ? |
This fixes everything! Thanks a lot for providing the fixes @larsoner I'm guessing these fixes will be available in 0.22.* release? |
thanks @larsoner ! feel free to backport this and make a minor release if you want |
Let's just be explicit about the encoding on disk everywhere by naming some constants for each reader.
Closes #8616
@Aniket-Pradhan can you see if this branch fixes the issues? FYI we're going to release 0.22 in the next couple of weeks if it matters for packaging purposes. so it might be worth