On Nov 1, 5:08 am, "Georg S. Weber" <[EMAIL PROTECTED]>
wrote:
> Hi,
Hi Georg,
> Sage3.2.alpha2 built fine on my Intel Core2 Duo Mac OS X 10.4.11 /
> Xcode 2.5.
Good.
> Due most probably to the patch from trac ticket #788, several doctests
> throw numerical noise errors.
Yeah, that was to be expected.
> Here are the first ones (doctesting is not complete yet):
>
> sage -t -long devel/sage/sage/calculus/calculus.py
> **********************************************************************
> File "/Users/georgweber/Public/sage/sage-3.2.alpha2/tmp/calculus.py",
> line 120:
> sage: float(f(pi))
> Expected:
> 6.1232339957367663e-16
> Got:
> 6.1230317691118863e-16
> **********************************************************************
> 1 items had failures:
> 1 of 76 in __main__.example_0
> ***Test Failed*** 1 failures.
> For whitespace errors, see the file /Users/georgweber/Public/sage/
> sage-3.2.alpha2/tmp/.doctest_calculus.py
> [76.7 s]
>
> sage -t -long devel/sage/sage/calculus/wester.py
> **********************************************************************
> File "/Users/georgweber/Public/sage/sage-3.2.alpha2/tmp/wester.py",
> line 261:
> : [float(f(i/10)) for i in range(1,5)]
> Expected:
> <BLANKLINE>
> [-0.00033670040754082975,
> -0.0027778004096620235,
> -0.0098909940914040928,
> -0.025411145508414501]
> Got:
> [-0.00033670040754082975, -0.0027778004096620235,
> -0.0098909940914040373, -0.025411145508414501]
> **********************************************************************
> 1 items had failures:
> 1 of 193 in __main__.example_0
> ***Test Failed*** 1 failures.
> For whitespace errors, see the file /Users/georgweber/Public/sage/
> sage-3.2.alpha2/tmp/.doctest_wester.py
> [13.9 s]
>
> sage -t -long devel/sage/sage/matrix/
> matrix_complex_double_dense.pyx********************************************
> **************************
> File "/Users/georgweber/Public/sage/sage-3.2.alpha2/tmp/
> matrix_complex_double_dense.py", line 899:
> sage: U*S*V.transpose()
> Expected:
> [...e-1... 1.0]
> [ 2.0 3.0]
> [ 4.0 5.0]
> Got:
> [ 0 1.0]
> [2.0 3.0]
> [4.0 5.0]
> **********************************************************************
> 1 items had failures:
> 1 of 18 in __main__.example_15
> ***Test Failed*** 1 failures.
> For whitespace errors, see the file /Users/georgweber/Public/sage/
> sage-3.2.alpha2/tmp/.doctest_matrix_complex_double_dense.py
> [4.3 s]
>
> sage -t -long devel/sage/sage/matrix/
> matrix_real_double_dense.pyx***********************************************
> ***********************
> File "/Users/georgweber/Public/sage/sage-3.2.alpha2/tmp/
> matrix_real_double_dense.py", line 786:
> sage: U*S*V.transpose()
> Expected:
> [...e-1... 1.0 2.0]
> [ 3.0 4.0 5.0]
> Got:
> [0.0 1.0 2.0]
> [3.0 4.0 5.0]
> **********************************************************************
> File "/Users/georgweber/Public/sage/sage-3.2.alpha2/tmp/
> matrix_real_double_dense.py", line 794:
> sage: U*S*V.transpose()
> Expected:
> [...e-1... 1.0]
> [ 2.0 3.0]
> [ 4.0 5.0]
> Got:
> [0.0 1.0]
> [2.0 3.0]
> [4.0 5.0]
> **********************************************************************
> 1 items had failures:
> 2 of 22 in __main__.example_16
> ***Test Failed*** 2 failures.
> For whitespace errors, see the file /Users/georgweber/Public/sage/
> sage-3.2.alpha2/tmp/.doctest_matrix_real_double_dense.py
> [3.4 s]
>
> I think it's strange that in the complex matrix case, the upper left
> entry is "0", and in the real matrix case it is "0.0".
> Should we have different trac tickets be opened for every file, or
> should we have one big "#788 follow-up" ticket?
I would make one ticket unless one particular file has a lot of
failures. I also build on a G5 and two Itaniums, so there might be
additional numerical noise problems.
But I would recommend that we make the failed tests in
matrix_real_double_dense.pyx and matrix_real_double_dense.pyx #random
again and add back the expected results with "0" in the upper left
corner. This one should definitely be a different ticket. Once we have
some infrastructure that asserts that a given vector/matrix is close
enough to the expected result we can change those doctests (there is
already a ticket).
> Cheers,
> gsw
Cheers,
Michael
--~--~---------~--~----~------------~-------~--~----~
To post to this group, send email to sage-devel@googlegroups.com
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at http://groups.google.com/group/sage-devel
URLs: http://www.sagemath.org
-~----------~----~----~----~------~----~------~--~---
