Buffer dtype mismatch, expected 'long' but got 'long long'

[malkwax]

hi, there is an error when I get nodal stress using NodalStress() method. I use the rst file in your closed issues "test_pyansys.rst".
I guess there is something wrong with numpy.int64 type.

the code is below:

stress = result.NodalStress(0)
Traceback (most recent call last):

  File "<ipython-input-12-6342d70f8bcd>", line 1, in <module>
    stress = result.NodalStress(0)

  File "C:\Users\zxmy\Anaconda3\lib\site-packages\pyansys\binary_reader.py", line 977, in NodalStress
    self.edge_idx.astype(np.int64))

  File "pyansys/cython/_rstHelper.pyx", line 158, in pyansys._rstHelper.LoadStress

ValueError: Buffer dtype mismatch, expected 'long' but got 'long long'

[akaszynski]

Didn't test the latest release on windows. I'll post an update shortly.

[malkwax]

thanks for your effort. I like it .

[akaszynski]

Fixed the code. Thanks for posting the bug reports!

[malkwax]

I install the update, but the bug is still there.
`stress=result.NodalStress(0)
Traceback (most recent call last):

File "", line 1, in
stress=result.NodalStress(0)

File "C:\Users\zxmy\Anaconda3\lib\site-packages\pyansys\binary_reader.py", line 979, in NodalStress
self.edge_idx.astype(c_long))

File "pyansys\cython_rstHelper.pyx", line 192, in pyansys._rstHelper.LoadStress

ValueError: Buffer dtype mismatch, expected 'long' but got 'long long'`

[akaszynski]

Try the new version 0.22.3 please. It’s here and on pypi.

[malkwax]

yes. it is perfect. thank you very much.

[akaszynski]

Ok. I'm closing this issue then. Thanks for testing it out. I'll be a bit more diligent about keeping cython platform independent. To anyone out there who's looking to make platform independent code when using numpy and cython, use the following:

from libc.stdint cimport int32_t, int64_t
# instead of int and long

Otherwise, you'll run into issues where an np.int64 is a long for linux and a long long for windows.

[jay-thakur]

Ok. I'm closing this issue then. Thanks for testing it out. I'll be a bit more diligent about keeping cython platform independent. To anyone out there who's looking to make platform independent code when using numpy and cython, use the following:

from libc.stdint cimport int32_t, int64_t
# instead of int and long

Otherwise, you'll run into issues where an np.int64 is a long for linux and a long long for windows.

Hi @akaszynski ,

I am facing similar issue while running a git code on windows, posted issue here . Could you please help me.