xtensor tests fail on less common architectures

xtensor tests fail on less common machine architectures, including i386, mipsel, s390x, and others. Build logs can be found at https://buildd.debian.org/status/package.php?p=xtensor&suite=experimental

There seem to be two separate sets of failures. test_xnpy fails on s390x, hppa, powerpc, ppc64, sparc64.

On i386 the failing tests are test_xbuilder, test_xfunction, test_xreducer, test_xnan_functions.

The xnpy error message on s390x is

[----------] 3 tests from xnpy
[ RUN      ] xnpy.load
unknown file: Failure
C++ exception with description "Cast error: formats not matching <f8 vs >f8" thrown in the test body.
[  FAILED  ] xnpy.load (3 ms)
[ RUN      ] xnpy.dump
./test/test_xnpy.cpp:146: Failure
Value of: compare_binary_files(filename, compare_name)
  Actual: false
Expected: true
./test/test_xnpy.cpp:150: Failure
Expected equality of these values:
  ularr_str
    Which is: "\x93NUMPY\x1\0v\0{'descr': '>u8', 'fortran_order': False, 'shape': (5,), }                                                            \n\0\0\0\0\0\0\0\f\0\0\0\0\0\0\0\xE\0\0\0\0\0\0\0\x10\0\0\0\0\0\0\0\x12\0\0\0\0\0\x12\xD5\x91"
  ularr_disk
    Which is: "\x93NUMPY\x1\0v\0{'descr': '<u8', 'fortran_order': False, 'shape': (5,), }                                                            \n\f\0\0\0\0\0\0\0\xE\0\0\0\0\0\0\0\x10\0\0\0\0\0\0\0\x12\0\0\0\0\0\0\0\x91\xD5\x12\0\0\0\0\0"
With diff:
@@ -1,2 +1,2 @@
-\x93NUMPY\x1\0v\0{'descr': '>u8', 'fortran_order': False, 'shape': (5,), }                                                            
-\0\0\0\0\0\0\0\f\0\0\0\0\0\0\0\xE\0\0\0\0\0\0\0\x10\0\0\0\0\0\0\0\x12\0\0\0\0\0\x12\xD5\x91
+\x93NUMPY\x1\0v\0{'descr': '<u8', 'fortran_order': False, 'shape': (5,), }                                                            
+\f\0\0\0\0\0\0\0\xE\0\0\0\0\0\0\0\x10\0\0\0\0\0\0\0\x12\0\0\0\0\0\0\0\x91\xD5\x12\0\0\0\0\0

Dumping boolean numpy file to string failed
[  FAILED  ] xnpy.dump (1 ms)
[ RUN      ] xnpy.xfunction_cast
unknown file: Failure
C++ exception with description "Cast error: formats not matching <f8 vs >f8" thrown in the test body.
[  FAILED  ] xnpy.xfunction_cast (0 ms)
[----------] 3 tests from xnpy (4 ms total)

The error messages on i386 are

[ RUN      ] xbuilder.linspace_double
./test/test_xbuilder.cpp:691: Failure
Expected equality of these values:
  a
    Which is: { 0, 2.04082, 4.08163, 6.12245, 8.16327, 10.2041, 12.2449, 14.2857, 16.3265, 18.3673, 20.4082, 22.449, 24.4898, 26.5306, 28.5714, 30.6122, 32.6531, 34.6939, 36.7347, 38.7755, 40.8163, 42.8571, 44.898, 46.9388, 48.9796, 51.0204, 53.0612, 55.102, 57.1429, 59.1837, 61.2245, 63.2653, ... }
  b
    Which is: { 0, 2.04082, 4.08163, 6.12245, 8.16327, 10.2041, 12.2449, 14.2857, 16.3265, 18.3673, 20.4082, 22.449, 24.4898, 26.5306, 28.5714, 30.6122, 32.6531, 34.6939, 36.7347, 38.7755, 40.8163, 42.8571, 44.898, 46.9388, 48.9796, 51.0204, 53.0612, 55.102, 57.1429, 59.1837, 61.2245, 63.2653, ... }
[  FAILED  ] xbuilder.linspace_double (0 ms)
[----------] 31 tests from xbuilder (1 ms total)

[----------] Global test environment tear-down
[==========] 31 tests from 1 test suite ran. (1 ms total)
[  PASSED  ] 30 tests.
[  FAILED  ] 1 test, listed below:
[  FAILED  ] xbuilder.linspace_double

[ RUN      ] xreducer.average
./test/test_xreducer.cpp:604: Failure
Value of: all(equal(avg1, expect1))
  Actual: false
Expected: true
./test/test_xreducer.cpp:605: Failure
Value of: all(equal(avg_m1, expect1))
  Actual: false
Expected: true
./test/test_xreducer.cpp:606: Failure
Value of: all(equal(avg_d1, expect1))
  Actual: false
Expected: true
[  FAILED  ] xreducer.average (10 ms)

[ RUN      ] xfunction.all_iterators
./test/test_xfunction.cpp:437: Failure
Expected equality of these values:
  res2
    Which is: { 0.036708, 1.39448, 1.87741, 0.0977121, 0.189934, 0.598901, 0.0709755, 1.97163, 1.98587, 1.87287, 1.05368, 1.92515, 0.0241392, 1.27395, 0.176453, 4.57634e-05, 0.0252965, 0.818594, 1.54328, 0.507352, 1.27338, 0.261067, 0.0898238, 0.928502, 0.317985, 1.09712, 0.450792, 0.356316, 0.0604582, 0.182303, 1.27131, 0.200408, ... }
  res5
    Which is: { 0.036708, 1.39448, 1.87741, 0.0977121, 0.189934, 0.598901, 0.0709755, 1.97163, 1.98587, 1.87287, 1.05368, 1.92515, 0.0241392, 1.27395, 0.176453, 4.57634e-05, 0.0252965, 0.818594, 1.54328, 0.507352, 1.27338, 0.261067, 0.0898238, 0.928502, 0.317985, 1.09712, 0.450792, 0.356316, 0.0604582, 0.182303, 1.27131, 0.200408, ... }
./test/test_xfunction.cpp:438: Failure
Expected equality of these values:
  res4
    Which is: { 0.036708, 1.39448, 1.87741, 0.0977121, 0.189934, 0.598901, 0.0709755, 1.97163, 1.98587, 1.87287, 1.05368, 1.92515, 0.0241392, 1.27395, 0.176453, 4.57634e-05, 0.0252965, 0.818594, 1.54328, 0.507352, 1.27338, 0.261067, 0.0898238, 0.928502, 0.317985, 1.09712, 0.450792, 0.356316, 0.0604582, 0.182303, 1.27131, 0.200408, ... }
  res5
    Which is: { 0.036708, 1.39448, 1.87741, 0.0977121, 0.189934, 0.598901, 0.0709755, 1.97163, 1.98587, 1.87287, 1.05368, 1.92515, 0.0241392, 1.27395, 0.176453, 4.57634e-05, 0.0252965, 0.818594, 1.54328, 0.507352, 1.27338, 0.261067, 0.0898238, 0.928502, 0.317985, 1.09712, 0.450792, 0.356316, 0.0604582, 0.182303, 1.27131, 0.200408, ... }
./test/test_xfunction.cpp:441: Failure
Expected equality of these values:
  res1
    Which is: { 0.036708, 1.39448, 1.87741, 0.0977121, 0.189934, 0.598901, 0.0709755, 1.97163, 1.98587, 1.87287, 1.05368, 1.92515, 0.0241392, 1.27395, 0.176453, 4.57634e-05, 0.0252965, 0.818594, 1.54328, 0.507352, 1.27338, 0.261067, 0.0898238, 0.928502, 0.317985, 1.09712, 0.450792, 0.356316, 0.0604582, 0.182303, 1.27131, 0.200408, ... }
  res5
    Which is: { 0.036708, 1.39448, 1.87741, 0.0977121, 0.189934, 0.598901, 0.0709755, 1.97163, 1.98587, 1.87287, 1.05368, 1.92515, 0.0241392, 1.27395, 0.176453, 4.57634e-05, 0.0252965, 0.818594, 1.54328, 0.507352, 1.27338, 0.261067, 0.0898238, 0.928502, 0.317985, 1.09712, 0.450792, 0.356316, 0.0604582, 0.182303, 1.27131, 0.200408, ... }
./test/test_xfunction.cpp:442: Failure
Expected equality of these values:
  res3
    Which is: { 0.036708, 1.39448, 1.87741, 0.0977121, 0.189934, 0.598901, 0.0709755, 1.97163, 1.98587, 1.87287, 1.05368, 1.92515, 0.0241392, 1.27395, 0.176453, 4.57634e-05, 0.0252965, 0.818594, 1.54328, 0.507352, 1.27338, 0.261067, 0.0898238, 0.928502, 0.317985, 1.09712, 0.450792, 0.356316, 0.0604582, 0.182303, 1.27131, 0.200408, ... }
  res5
    Which is: { 0.036708, 1.39448, 1.87741, 0.0977121, 0.189934, 0.598901, 0.0709755, 1.97163, 1.98587, 1.87287, 1.05368, 1.92515, 0.0241392, 1.27395, 0.176453, 4.57634e-05, 0.0252965, 0.818594, 1.54328, 0.507352, 1.27338, 0.261067, 0.0898238, 0.928502, 0.317985, 1.09712, 0.450792, 0.356316, 0.0604582, 0.182303, 1.27131, 0.200408, ... }
./test/test_xfunction.cpp:443: Failure
Expected equality of these values:
  func
    Which is: { 0.036708, 1.39448, 1.87741, 0.0977121, 0.189934, 0.598901, 0.0709755, 1.97163, 1.98587, 1.87287, 1.05368, 1.92515, 0.0241392, 1.27395, 0.176453, 4.57634e-05, 0.0252965, 0.818594, 1.54328, 0.507352, 1.27338, 0.261067, 0.0898238, 0.928502, 0.317985, 1.09712, 0.450792, 0.356316, 0.0604582, 0.182303, 1.27131, 0.200408, ... }
  res5
    Which is: { 0.036708, 1.39448, 1.87741, 0.0977121, 0.189934, 0.598901, 0.0709755, 1.97163, 1.98587, 1.87287, 1.05368, 1.92515, 0.0241392, 1.27395, 0.176453, 4.57634e-05, 0.0252965, 0.818594, 1.54328, 0.507352, 1.27338, 0.261067, 0.0898238, 0.928502, 0.317985, 1.09712, 0.450792, 0.356316, 0.0604582, 0.182303, 1.27131, 0.200408, ... }
./test/test_xfunction.cpp:441: Failure
Expected equality of these values:
  res1
    Which is: { 0.276178, 0.867324, 2.48769, 0.0298791, 0.610281, 1.11554, 0.195672, 2.54935, 0.1347, 1.91638, 1.47967, 1.01908, 0.282084, 0.107887, 0.588225, 0.00975142, 0.116817, 1.64267, 0.118745, 0.997433, 1.62662, 0.375276, 0.544142, 0.0921053, 0.789645, 1.50985, 0.493132, 1.08376, 0.0235029, 0.597896, 1.6253, 0.328801, ... }
  res5
    Which is: { 0.276178, 0.867324, 2.48769, 0.0298791, 0.610281, 1.11554, 0.195672, 2.54935, 0.1347, 1.91638, 1.47967, 1.01908, 0.282084, 0.107887, 0.588225, 0.00975142, 0.116817, 1.64267, 0.118745, 0.997433, 1.62662, 0.375276, 0.544142, 0.0921053, 0.789645, 1.50985, 0.493132, 1.08376, 0.0235029, 0.597896, 1.6253, 0.328801, ... }
./test/test_xfunction.cpp:442: Failure
Expected equality of these values:
  res3
    Which is: { 0.276178, 0.867324, 2.48769, 0.0298791, 0.610281, 1.11554, 0.195672, 2.54935, 0.1347, 1.91638, 1.47967, 1.01908, 0.282084, 0.107887, 0.588225, 0.00975142, 0.116817, 1.64267, 0.118745, 0.997433, 1.62662, 0.375276, 0.544142, 0.0921053, 0.789645, 1.50985, 0.493132, 1.08376, 0.0235029, 0.597896, 1.6253, 0.328801, ... }
  res5
    Which is: { 0.276178, 0.867324, 2.48769, 0.0298791, 0.610281, 1.11554, 0.195672, 2.54935, 0.1347, 1.91638, 1.47967, 1.01908, 0.282084, 0.107887, 0.588225, 0.00975142, 0.116817, 1.64267, 0.118745, 0.997433, 1.62662, 0.375276, 0.544142, 0.0921053, 0.789645, 1.50985, 0.493132, 1.08376, 0.0235029, 0.597896, 1.6253, 0.328801, ... }
./test/test_xfunction.cpp:443: Failure
Expected equality of these values:
  func
    Which is: { 0.276178, 0.867324, 2.48769, 0.0298791, 0.610281, 1.11554, 0.195672, 2.54935, 0.1347, 1.91638, 1.47967, 1.01908, 0.282084, 0.107887, 0.588225, 0.00975142, 0.116817, 1.64267, 0.118745, 0.997433, 1.62662, 0.375276, 0.544142, 0.0921053, 0.789645, 1.50985, 0.493132, 1.08376, 0.0235029, 0.597896, 1.6253, 0.328801, ... }
  res5
    Which is: { 0.276178, 0.867324, 2.48769, 0.0298791, 0.610281, 1.11554, 0.195672, 2.54935, 0.1347, 1.91638, 1.47967, 1.01908, 0.282084, 0.107887, 0.588225, 0.00975142, 0.116817, 1.64267, 0.118745, 0.997433, 1.62662, 0.375276, 0.544142, 0.0921053, 0.789645, 1.50985, 0.493132, 1.08376, 0.0235029, 0.597896, 1.6253, 0.328801, ... }
./test/test_xfunction.cpp:441: Failure
Expected equality of these values:
  res1
    Which is: { 5.27618, 5.86732, 7.48769, 5.02988, 5.61028, 6.11554, 5.19567, 7.54935, 5.1347, 6.91638, 6.47967, 6.01908, 5.28208, 5.10789, 5.58822, 5.00975, 5.11682, 6.64267, 5.11875, 5.99743, 6.62662, 5.37528, 5.54414, 5.09211, 5.78965, 6.50985, 5.49313, 6.08376, 5.0235, 5.5979, 6.6253, 5.3288, ... }
  res5
    Which is: { 5.27618, 5.86732, 7.48769, 5.02988, 5.61028, 6.11554, 5.19567, 7.54935, 5.1347, 6.91638, 6.47967, 6.01908, 5.28208, 5.10789, 5.58822, 5.00975, 5.11682, 6.64267, 5.11875, 5.99743, 6.62662, 5.37528, 5.54414, 5.09211, 5.78965, 6.50985, 5.49313, 6.08376, 5.0235, 5.5979, 6.6253, 5.3288, ... }
./test/test_xfunction.cpp:442: Failure
Expected equality of these values:
  res3
    Which is: { 5.27618, 5.86732, 7.48769, 5.02988, 5.61028, 6.11554, 5.19567, 7.54935, 5.1347, 6.91638, 6.47967, 6.01908, 5.28208, 5.10789, 5.58822, 5.00975, 5.11682, 6.64267, 5.11875, 5.99743, 6.62662, 5.37528, 5.54414, 5.09211, 5.78965, 6.50985, 5.49313, 6.08376, 5.0235, 5.5979, 6.6253, 5.3288, ... }
  res5
    Which is: { 5.27618, 5.86732, 7.48769, 5.02988, 5.61028, 6.11554, 5.19567, 7.54935, 5.1347, 6.91638, 6.47967, 6.01908, 5.28208, 5.10789, 5.58822, 5.00975, 5.11682, 6.64267, 5.11875, 5.99743, 6.62662, 5.37528, 5.54414, 5.09211, 5.78965, 6.50985, 5.49313, 6.08376, 5.0235, 5.5979, 6.6253, 5.3288, ... }
./test/test_xfunction.cpp:443: Failure
Expected equality of these values:
  func
    Which is: { 5.27618, 5.86732, 7.48769, 5.02988, 5.61028, 6.11554, 5.19567, 7.54935, 5.1347, 6.91638, 6.47967, 6.01908, 5.28208, 5.10789, 5.58822, 5.00975, 5.11682, 6.64267, 5.11875, 5.99743, 6.62662, 5.37528, 5.54414, 5.09211, 5.78965, 6.50985, 5.49313, 6.08376, 5.0235, 5.5979, 6.6253, 5.3288, ... }
  res5
    Which is: { 5.27618, 5.86732, 7.48769, 5.02988, 5.61028, 6.11554, 5.19567, 7.54935, 5.1347, 6.91638, 6.47967, 6.01908, 5.28208, 5.10789, 5.58822, 5.00975, 5.11682, 6.64267, 5.11875, 5.99743, 6.62662, 5.37528, 5.54414, 5.09211, 5.78965, 6.50985, 5.49313, 6.08376, 5.0235, 5.5979, 6.6253, 5.3288, ... }
./test/test_xfunction.cpp:441: Failure
Expected equality of these values:
  res1
    Which is: { 0.069039, 0.0384502, 0.814603, 0.000949708, 0.0575034, 0.298263, 0.00505859, 0.53314, 0.000363417, 0.434563, 0.263347, 0.149012, 0.0187634, 0.00496284, 0.0749659, 0.00247823, 0.0015759, 0.138508, 0.0199797, 0.273498, 0.532298, 0.103008, 0.0547457, 0.00147197, 0.0652372, 0.377434, 0.0218615, 0.354882, 0.00074704, 0.0563364, 0.434558, 0.00850028, ... }
  res5
    Which is: { 0.069039, 0.0384502, 0.814603, 0.000949708, 0.0575034, 0.298263, 0.00505859, 0.53314, 0.000363417, 0.434563, 0.263347, 0.149012, 0.0187634, 0.00496284, 0.0749659, 0.00247823, 0.0015759, 0.138508, 0.0199797, 0.273498, 0.532298, 0.103008, 0.0547457, 0.00147197, 0.0652372, 0.377434, 0.0218615, 0.354882, 0.00074704, 0.0563364, 0.434558, 0.00850028, ... }
./test/test_xfunction.cpp:442: Failure
Expected equality of these values:
  res3
    Which is: { 0.069039, 0.0384502, 0.814603, 0.000949708, 0.0575034, 0.298263, 0.00505859, 0.53314, 0.000363417, 0.434563, 0.263347, 0.149012, 0.0187634, 0.00496284, 0.0749659, 0.00247823, 0.0015759, 0.138508, 0.0199797, 0.273498, 0.532298, 0.103008, 0.0547457, 0.00147197, 0.0652372, 0.377434, 0.0218615, 0.354882, 0.00074704, 0.0563364, 0.434558, 0.00850028, ... }
  res5
    Which is: { 0.069039, 0.0384502, 0.814603, 0.000949708, 0.0575034, 0.298263, 0.00505859, 0.53314, 0.000363417, 0.434563, 0.263347, 0.149012, 0.0187634, 0.00496284, 0.0749659, 0.00247823, 0.0015759, 0.138508, 0.0199797, 0.273498, 0.532298, 0.103008, 0.0547457, 0.00147197, 0.0652372, 0.377434, 0.0218615, 0.354882, 0.00074704, 0.0563364, 0.434558, 0.00850028, ... }
./test/test_xfunction.cpp:443: Failure
Expected equality of these values:
  func
    Which is: { 0.069039, 0.0384502, 0.814603, 0.000949708, 0.0575034, 0.298263, 0.00505859, 0.53314, 0.000363417, 0.434563, 0.263347, 0.149012, 0.0187634, 0.00496284, 0.0749659, 0.00247823, 0.0015759, 0.138508, 0.0199797, 0.273498, 0.532298, 0.103008, 0.0547457, 0.00147197, 0.0652372, 0.377434, 0.0218615, 0.354882, 0.00074704, 0.0563364, 0.434558, 0.00850028, ... }
  res5
    Which is: { 0.069039, 0.0384502, 0.814603, 0.000949708, 0.0575034, 0.298263, 0.00505859, 0.53314, 0.000363417, 0.434563, 0.263347, 0.149012, 0.0187634, 0.00496284, 0.0749659, 0.00247823, 0.0015759, 0.138508, 0.0199797, 0.273498, 0.532298, 0.103008, 0.0547457, 0.00147197, 0.0652372, 0.377434, 0.0218615, 0.354882, 0.00074704, 0.0563364, 0.434558, 0.00850028, ... }
./test/test_xfunction.cpp:441: Failure
Expected equality of these values:
  res1
    Which is: { 0.26585, 18.1336, 1.15235, 51.4432, 1.6515, 1.00398, 7.01535, 1.84907, 2732.22, 2.15488, 2.00056, 6.45972, 0.643253, 128.348, 1.17689, 0.00923309, 8.02604, 2.95504, 38.6212, 0.927526, 1.19612, 1.26722, 0.820374, 315.393, 2.43714, 1.45339, 10.3102, 0.50202, 40.4652, 1.61799, 1.46276, 11.7883, ... }
  res5
    Which is: { 0.26585, 18.1336, 1.15235, 51.4432, 1.6515, 1.00398, 7.01535, 1.84907, 2732.22, 2.15488, 2.00056, 6.45972, 0.643253, 128.348, 1.17689, 0.00923309, 8.02604, 2.95504, 38.6212, 0.927526, 1.19612, 1.26722, 0.820374, 315.393, 2.43714, 1.45339, 10.3102, 0.50202, 40.4652, 1.61799, 1.46276, 11.7883, ... }
./test/test_xfunction.cpp:442: Failure
Expected equality of these values:
  res3
    Which is: { 0.26585, 18.1336, 1.15235, 51.4432, 1.6515, 1.00398, 7.01535, 1.84907, 2732.22, 2.15488, 2.00056, 6.45972, 0.643253, 128.348, 1.17689, 0.00923309, 8.02604, 2.95504, 38.6212, 0.927526, 1.19612, 1.26722, 0.820374, 315.393, 2.43714, 1.45339, 10.3102, 0.50202, 40.4652, 1.61799, 1.46276, 11.7883, ... }
  res5
    Which is: { 0.26585, 18.1336, 1.15235, 51.4432, 1.6515, 1.00398, 7.01535, 1.84907, 2732.22, 2.15488, 2.00056, 6.45972, 0.643253, 128.348, 1.17689, 0.00923309, 8.02604, 2.95504, 38.6212, 0.927526, 1.19612, 1.26722, 0.820374, 315.393, 2.43714, 1.45339, 10.3102, 0.50202, 40.4652, 1.61799, 1.46276, 11.7883, ... }
./test/test_xfunction.cpp:443: Failure
Expected equality of these values:
  func
    Which is: { 0.26585, 18.1336, 1.15235, 51.4432, 1.6515, 1.00398, 7.01535, 1.84907, 2732.22, 2.15488, 2.00056, 6.45972, 0.643253, 128.348, 1.17689, 0.00923309, 8.02604, 2.95504, 38.6212, 0.927526, 1.19612, 1.26722, 0.820374, 315.393, 2.43714, 1.45339, 10.3102, 0.50202, 40.4652, 1.61799, 1.46276, 11.7883, ... }
  res5
    Which is: { 0.26585, 18.1336, 1.15235, 51.4432, 1.6515, 1.00398, 7.01535, 1.84907, 2732.22, 2.15488, 2.00056, 6.45972, 0.643253, 128.348, 1.17689, 0.00923309, 8.02604, 2.95504, 38.6212, 0.927526, 1.19612, 1.26722, 0.820374, 315.393, 2.43714, 1.45339, 10.3102, 0.50202, 40.4652, 1.61799, 1.46276, 11.7883, ... }
[  FAILED  ] xfunction.all_iterators (6 ms)

[ RUN      ] xnanfunctions.nanmean
./test/test_xnan_functions.cpp:157: Failure
Expected equality of these values:
  nanmean(nantest::aN, {1})
    Which is: { 63, 2, 1.66667 }
  eaN1
    Which is: { 63, 2, 1.66667 }
./test/test_xnan_functions.cpp:163: Failure
Expected equality of these values:
  nanmean(nantest::aN, {1}, evaluation_strategy::immediate)
    Which is: { 63, 2, 1.66667 }
  eaN1
    Which is: { 63, 2, 1.66667 }
./test/test_xnan_functions.cpp:177: Failure
Expected equality of these values:
  nanmean(nantest::cN, {1}, evaluation_strategy::immediate)
    Which is: { (1,1), (1.66667,0.333333) }
  ecN1
    Which is: { (1,1), (1.66667,0.333333) }
[  FAILED  ] xnanfunctions.nanmean (0 ms)

Mar 17 '21 10:03 drew-parsons

Do you have the same error on all architectures for test_xnpy? From the message you've posted, it is definitely a big endian / little endian issue (we need to fix the expected value in the test since we always assume little endian). Regarding the other failures, they look like precision / rounding issues.

Is it easy for you to test any branch from Github? If so, I can push some changes that will help to diagnose the problem.

Mar 23 '21 13:03 JohanMabille

Not all arches. The test_xnpy failure is on s390x, hppa, powerpc, ppc64, sparc64. It passes on ARM in both armel (little endian) and armhf (arm64 also passes).

I can pull branches if needed to test (I can test on Debian experimental so the main distribution is not affected).

Mar 23 '21 15:03 drew-parsons

That is, test_xnpy passed on amd64, arm64, armel, armhf, i386, mips64el, ppc64el, alpha, ia64, m68k, riscv64, sh4, x32.

Mar 23 '21 20:03 drew-parsons

With xtensor 0.23.10, other tests now pass. Only the 3 tests in test_xnpy continue to fail. For s390x:

[----------] 3 tests from xnpy
[ RUN      ] xnpy.load
unknown file: Failure
C++ exception with description "Cast error: formats not matching <f8 vs >f8" thrown in the test body.
[  FAILED  ] xnpy.load (7 ms)
[ RUN      ] xnpy.dump
./test/test_xnpy.cpp:146: Failure
Value of: compare_binary_files(filename, compare_name)
  Actual: false
Expected: true
./test/test_xnpy.cpp:150: Failure
Expected equality of these values:
  ularr_str
    Which is: "\x93NUMPY\x1\0v\0{'descr': '>u8', 'fortran_order': False, 'shape': (5,), }                                                            \n\0\0\0\0\0\0\0\f\0\0\0\0\0\0\0\xE\0\0\0\0\0\0\0\x10\0\0\0\0\0\0\0\x12\0\0\0\0\0\x12\xD5\x91"
  ularr_disk
    Which is: "\x93NUMPY\x1\0v\0{'descr': '<u8', 'fortran_order': False, 'shape': (5,), }                                                            \n\f\0\0\0\0\0\0\0\xE\0\0\0\0\0\0\0\x10\0\0\0\0\0\0\0\x12\0\0\0\0\0\0\0\x91\xD5\x12\0\0\0\0\0"
With diff:
@@ -1,2 +1,2 @@
-\x93NUMPY\x1\0v\0{'descr': '>u8', 'fortran_order': False, 'shape': (5,), }                                                            
-\0\0\0\0\0\0\0\f\0\0\0\0\0\0\0\xE\0\0\0\0\0\0\0\x10\0\0\0\0\0\0\0\x12\0\0\0\0\0\x12\xD5\x91
+\x93NUMPY\x1\0v\0{'descr': '<u8', 'fortran_order': False, 'shape': (5,), }                                                            
+\f\0\0\0\0\0\0\0\xE\0\0\0\0\0\0\0\x10\0\0\0\0\0\0\0\x12\0\0\0\0\0\0\0\x91\xD5\x12\0\0\0\0\0

Dumping boolean numpy file to string failed
[  FAILED  ] xnpy.dump (2 ms)
[ RUN      ] xnpy.xfunction_cast
unknown file: Failure
C++ exception with description "Cast error: formats not matching <f8 vs >f8" thrown in the test body.
[  FAILED  ] xnpy.xfunction_cast (0 ms)
[----------] 3 tests from xnpy (9 ms total)

Sep 13 '21 14:09 drew-parsons

Hi, I see the same issue with 0.24.3 on ppc64 (big endian architecture).

[  955s] 60/79 Test #60: test_xnpy .........................***Failed    0.01 sec
[  955s] [doctest] doctest version is "2.4.8"
[  955s] [doctest] run with "--help" for options
[  955s] ===============================================================================
[  955s] /home/abuild/rpmbuild/BUILD/xtensor-0.24.3/test/test_xnpy.cpp:39:
[  955s] TEST CASE:  xnpy.load
[  955s] 
[  955s] /home/abuild/rpmbuild/BUILD/xtensor-0.24.3/test/test_xnpy.cpp:39: ERROR: test case THREW exception: Cast error: formats not matching <f8 vs >f8
[  955s] 
[  955s] ===============================================================================
[  955s] /home/abuild/rpmbuild/BUILD/xtensor-0.24.3/test/test_xnpy.cpp:111:
[  955s] TEST CASE:  xnpy.dump
[  955s] 
[  955s] /home/abuild/rpmbuild/BUILD/xtensor-0.24.3/test/test_xnpy.cpp:144: ERROR: CHECK_EQ( compare_binary_files(filename, compare_name), true ) is NOT correct!
[  955s]   values: CHECK_EQ( false, true )
[  955s] 
[  955s] /home/abuild/rpmbuild/BUILD/xtensor-0.24.3/test/test_xnpy.cpp:148: ERROR: CHECK( ularr_str==ularr_disk ) is NOT correct!
[  955s]   values: CHECK( NUMPY )
[  955s]   logged: Dumping boolean numpy file to string failed
[  955s] 
[  955s] ===============================================================================
[  955s] /home/abuild/rpmbuild/BUILD/xtensor-0.24.3/test/test_xnpy.cpp:153:
[  955s] TEST CASE:  xnpy.xfunction_cast
[  955s] 
[  955s] /home/abuild/rpmbuild/BUILD/xtensor-0.24.3/test/test_xnpy.cpp:153: ERROR: test case THREW exception: Cast error: formats not matching <f8 vs >f8
[  955s] 
[  955s] ===============================================================================
[  955s] [doctest] test cases: 3 | 0 passed | 3 failed | 0 skipped
[  955s] [doctest] assertions: 5 | 3 passed | 2 failed |
[  955s] [doctest] Status: FAILURE!
[  955s]

Oct 19 '22 09:10 matwey

The following is currently still failing for x86 arch for 0.24.7:

[  933s] [doctest] doctest version is "2.4.11"
[  933s] [doctest] run with "--help" for options
[  933s] Built without XSIMD
[  933s] Built without XSIMD
[  933s] ===============================================================================
[  933s] /home/abuild/rpmbuild/BUILD/xtensor-0.24.7/test/test_xbuilder.cpp:146:
[  933s] TEST CASE:  xbuilder.arange_min_max_step
[  933s] 
[  933s] /home/abuild/rpmbuild/BUILD/xtensor-0.24.7/test/test_xbuilder.cpp:166: FATAL ERROR: REQUIRE_EQ( 3.f * 0.3f, l3[{3}] ) is NOT correct!
[  933s]   values: REQUIRE_EQ( 0.9, 0.9 )
[  933s] 
[  933s] ===============================================================================
[  933s] /home/abuild/rpmbuild/BUILD/xtensor-0.24.7/test/test_xcomplex.cpp:297:
[  933s] TEST CASE:  xcomplex.longdouble
[  933s] 
[  933s] /home/abuild/rpmbuild/BUILD/xtensor-0.24.7/test/test_xcomplex.cpp:304: ERROR: CHECK_EQ( a(4, 4), cmplx(123.321, -123.321) ) is NOT correct!
[  933s]   values: CHECK_EQ( (123.321,-123.321), (123.321,-123.321) )
[  933s] 
[  933s] /home/abuild/rpmbuild/BUILD/xtensor-0.24.7/test/test_xcomplex.cpp:308: ERROR: CHECK_EQ( a(0, 0), cmplx(-123.321, -123.321) ) is NOT correct!
[  933s]   values: CHECK_EQ( (-123.321,-123.321), (-123.321,-123.321) )
[  933s] 
[  933s] /home/abuild/rpmbuild/BUILD/xtensor-0.24.7/test/test_xcomplex.cpp:309: ERROR: CHECK_EQ( a(4, 4), cmplx(-123.321, -123.321) ) is NOT correct!
[  933s]   values: CHECK_EQ( (-123.321,-123.321), (-123.321,-123.321) )
[  933s] 
[  933s] ===============================================================================

Jan 04 '24 15:01 matwey

Not sure how to reimplement test_xbuilder.cpp. The failure appears when gcc is used (it seems to use FPU instead of SSE for floating point).

Jan 04 '24 20:01 matwey

Maybe we can use floating point comparisons / assertions instead

Jan 04 '24 20:01 JohanMabille

Maybe we can use floating point comparisons / assertions instead

I think it is enough to make the step be power of 2, like 0.25f

Jan 04 '24 20:01 matwey