Hi, On Mon, Jun 14, 2010 at 03:13:53PM -0700, daveloeffler wrote: > I've been doing some work adding doctests to some files that don't > have them, and the coverage script kept telling me to put in a > "TestSuite.run()" doctest. > > Can someone tell me how on earth I can get one of these to actually > work? No matter what I do, if X is a parent object, > "TestSuite(X).run()" always fails at the point where it creates an > element of X and runs "_test_category" on it.
1 - Did you properly initialize category for your parent ? Looking at the code from rings/ideal_monoid.py there isn't any category specified for the constructor of Ideal_monoid_c. 2 - Did you specify in the parent how the element are represented ? There isn't any definition of Element. 3 - If you don't want your object to get category stuff, you can simply deactivate this test using the option skip=... > It seems to want > elements of X to be instances of a "dynamic metaclass" created on the > fly by the categories machinery. How is that supposed to work? How are > rich, complicated objects (e.g. ideals in arbitrary rings -- I was > working on rings/ideal_monoid.py) supposed to be represented by a > class that's cooked up out of nowhere? I don't see the relevance of having rich objects in the discussion. > I looked for examples where TestSuite doctests already exist (and > pass), but in all the ones I found either the element class didn't > really do anything (e.g. the examples in sage/categories/examples) or > the element class was a Cython extension type and thus exempt from > this test. The idea (maybe it is not a good idea) of the doc in sage/categories/examples is to provide minimal example. That why most of the things are trivial here. Still the file sage/categories/examples/sets_cat.py demonstrate various way of creating representing elements of your class: - either by inheriting from an already written class - or by wrapping element of an already build class - or by having a facade parent that is a parent which deals with elements whose parent is actually another parent (e.g: Prime v.s. Integer) Cheers, Florent -- 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
