On Mon, Mar 09, 2020 at 08:25:03AM +0100, Martin Reindl wrote:
> On Mon, Mar 09, 2020 at 08:07:23AM +0100, Theo Buehler wrote:
> > > Is this the complete error output? I get the impression that errors
> > > related to
> > > numpy are common to hdf5/netcdf ports, it's probably similiar to how
> > > py-netcdf
> > > fails.
> >
> > Yes, it's the complete output inlined from test.log. I'm building
> > math/py-netcdf4 now which looks like it will take a while, will report
> > back once I will have run its tests.
> >
> > See also ~tb/py-h5py-test.log on cvs.
>
> If you can spare the CPU cycles, try with the numpy 1.16.5 diff floating
> around.
The py-netcdf4 failure (with the in-tree numpy) is an alignment issue.
Full test.log:
......................................../usr/ports/pobj/py-netcdf4-1.5.3/netCDF4-1.5.3/test/tst_types.py:92:
UserWarning: WARNING: missing_value not used since it
cannot be safely cast to variable data type
assert_array_equal(v3[:],-1*np.ones(n2dim,v3.dtype))
..................................................F
======================================================================
FAIL: runTest (tst_compoundvar.VariablesTestCase)
testing compound variables
----------------------------------------------------------------------
Traceback (most recent call last):
File
"/usr/ports/pobj/py-netcdf4-1.5.3/netCDF4-1.5.3/test/tst_compoundvar.py", line
72, in setUp
assert (cmptype4 == dtype4a) # data type should be aligned
AssertionError
----------------------------------------------------------------------
Ran 91 tests in 412.802s
FAILED (failures=1)
not running tst_unicode.py ...
netcdf4-python version: 1.5.3
HDF5 lib version: 1.10.6
netcdf lib version: 4.7.3
numpy version 1.14.6
cython version 0.29.14
foo_bar
*** Error 1 in . (Makefile:50 'do-test': @cd
/usr/ports/pobj/py-netcdf4-1.5.3/netCDF4-1.5.3/test && /usr/bin/env -i CC=cc
PYTHONUSERBASE= PO...)
