Am 17.07.19 um 07:25 schrieb Martin Reindl: > On Wed, Jul 17, 2019 at 12:50:34AM -0400, Kurt Mosiejczuk wrote: >> On Thu, Jun 27, 2019 at 03:14:31PM +0200, Martin Reindl wrote: >> On sparc64, py-netcdf4 fails two tests: test_unsigned and VariablesTestCase. >> >> ====================================================================== >> >> FAIL: test_unsigned (tst_Unsigned.Test_Unsigned) >> >> ---------------------------------------------------------------------- >> >> Traceback (most recent call last): >> File >> "/usr/ports/pobj/py-netcdf4-1.5.1.2-python3/netCDF4-1.5.1.2/test/tst_Unsi >> gned.py", line 31, in test_unsigned >> assert(data1.mask.sum() == data2.mask.sum()) >> AssertionError >> >> ====================================================================== >> FAIL: runTest (tst_compoundvar.VariablesTestCase) >> testing compound variables >> ---------------------------------------------------------------------- >> Traceback (most recent call last): >> File >> "/usr/ports/pobj/py-netcdf4-1.5.1.2-python3/netCDF4-1.5.1.2/test/tst_comp >> oundvar.py", line 72, in setUp >> assert (cmptype4 == dtype4a) # data type should be aligned >> AssertionError >> >> I don't know if it is their test code or some other difference on sparc64. >> > > Very grateful for your thorough testing and input for these ports. > > From a first glance this is a netcdf-specific code problem on sparc64. dtype4a > is numpy code where I suspect there is no problem on sparc64. Whereas cmdtype4 > uses netcdf's createCompoundType method. Need to look deeper into this as time > permits, even though I am not getting a sparc64 (again).
Picking this up again hoping to move forward with this. Back in July I traced this down to which seems to be a bug in math/hdf5 on sparc64. But I did not dig deeper. Does this justify marking py-netcdf BROKEN on sparc64 for the purpose of moving on? In the meantime both ports got updated (attached). Please review and comment. -m
py-netcdf153.tgz
Description: GNU Zip compressed data
py-cftime1042.tgz
Description: GNU Zip compressed data
