On 5/3/11 7:13 AM, Rob Beezer wrote:
And in this case such a doctest belongs somewhere else and the one "originating" doctest should be modified or permanently removed.
+1. The test should be in the SVD routine. In fact, numpy may want to add a failing doctest to their SVD routine. Since we call numpy to do the SVD, maybe we should just make sure a failing doctest gets added to numpy and call it good, since we run the 3rd party tests with spkg -check or something like that.
The general point still stands of the value of having known failures in doctests, even if this failing doctest is better off in a 3rd party library that stands between us and the bug; a 3rd party library that does have the concept of failing doctests.
Jason -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
