Ensure native byte order for memmap.

[cmarmo]

This pull request ensures native byte order for numpy memmap ndarray, generalizing the fix for ndarray applied in #1181.
It is supposed to fix #1353.

Thanks for considering it.

[codecov]

Codecov Report

Base: 93.93% // Head: 93.86% // Decreases project coverage by -0.06% ⚠️

Coverage data is based on head (e3cb5f3) compared to base (5d0cf09).
Patch coverage: 100.00% of modified lines in pull request are covered.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1354      +/-   ##
==========================================
- Coverage   93.93%   93.86%   -0.07%     
==========================================
  Files          52       52              
  Lines        7292     7292              
==========================================
- Hits         6850     6845       -5     
- Misses        442      447       +5     

Impacted Files	Coverage Δ
joblib/numpy_pickle.py	99.14% <100.00%> (ø)
joblib/test/test_store_backends.py	91.42% <0.00%> (-5.72%)	⬇️
joblib/_parallel_backends.py	91.54% <0.00%> (-1.48%)	⬇️
joblib/parallel.py	96.56% <0.00%> (+0.26%)	⬆️

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

☔ View full report at Codecov.
📢 Do you have feedback about the report comment? Let us know in this issue.

[ogrisel]

Thanks for the PR! This intuitively looks good to me but I don't have access to a big endian platform. Have you had the opportunity to run the tests against your branch on a big endian machine yourself?

Even if not, +1 for merging. Please document the change in the changelog targeting the next version.

[cmarmo]

Thanks for the PR! This intuitively looks good to me but I don't have access to a big endian platform. Have you had the opportunity to run the tests against your branch on a big endian machine yourself?

Unfortunately not, let me see with debian people if it is possible to test this on experimental...

Even if not, +1 for merging. Please document the change in the changelog targeting the next version.

Will do.

[cmarmo]

Thanks for the PR! This intuitively looks good to me but I don't have access to a big endian platform. Have you had the opportunity to run the tests against your branch on a big endian machine yourself?

Unfortunately not, let me see with debian people if it is possible to test this on experimental...

I can have this tested on debian machines but they are under a big software update right now so it will take some time... Iĺl let you know when it's done.

[nileshpatra]

Hi @ogrisel @cmarmo I tested this patch on a s390x machine (big endian) and it gets the test passing there.

[ogrisel]

Thanks very much to both of you!