FIX OSX test failures in qz decomposition. #292

Merged
merged 1 commit into from Sep 13, 2012

Projects

None yet

2 participants

@fabianp
Member
fabianp commented Aug 26, 2012

Test the absolut because the sign might be platform dependent. Tested
in OS X 10.7 x86-64.

@fabianp fabianp FIX OSX test failures in qz decomposition.
Test the absolut because the sign might be platform dependent. Tested
in OS X 10.7 x86-64.
7b73a57
@fabianp
Member
fabianp commented Aug 26, 2012

BTW @jseabold , I think testing

        assert_array_almost_equal(dot(dot(Q,AA),Z.T), A)
        assert_array_almost_equal(dot(dot(Q,BB),Z.T), B)

plus properties of Q, AA, BB would be enough to test the result is correct, and probably more readable than hard-coding the result, what do you think ?

@rgommers
Member
rgommers commented Sep 8, 2012

See #303, which disables the sort kw. I think we can merge this first and then #303. Please have a look, since this is still a blocker for the 0.11.0 release.

@rgommers
Member

OK merging this now; will backport to 0.11.x. The question of hardcoding values in the test can be revisited once we re-enable this functionality (disabled by #303). Thanks Fabian.

@rgommers rgommers merged commit 78fa7a2 into scipy:master Sep 13, 2012
@fabianp
Member
fabianp commented Sep 13, 2012

Ok thanks for taking care of this
On Sep 13, 2012 10:37 PM, "Ralf Gommers" notifications@github.com wrote:

OK merging this now; will backport to 0.11.x. The question of hardcoding
values in the test can be revisited once we re-enable this functionality
(disabled by #303 #303). Thanks
Fabian.


Reply to this email directly or view it on GitHubhttps://github.com/scipy/scipy/pull/292#issuecomment-8542494.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment