On Mon, May 10, 2010 at 12:08 PM, Christian Muise <christian.mu...@gmail.com> wrote: > Yowzas... > - Why the changes from "self.new(...)" to "self.__class__(...)"? (it happens > in a number of places)
Good point --- I removed .new(), as it used the old assumptions, but I might as well just introduce it back and call __class__ from there. > - I never know about facts.py :p. I was just trying the disconnect by > removing the inheritance from Basic(AssumMeths) > - Is there a canonical way I can search through all the outstanding issues > with the old assumptions deletion? Yes, run the tests in my branch and see the failures. > - You added back all of the is_* as properties (that trivially return False > for now). Shouldn't those eventually be stripped out and only the new > assumption system used? (via refine and ask) Maybe, I don't have an opinion on it as of now. The way I view it is that the new way in the assum branch is that "is_*" properties are just convenient functions (resp. properties not to break compatibility) that return you some useful information about the expression. It has nothing to do with assumptions. E.g. S(5).is_positive returns True, S(5).is_negative returns False and so on. Lots of code in the core depends on this and I don't see a reason why we should delete this. However, it should not use assumptions in any way (and it doesn't in my assum branch). > - Finally, should I be picking up on any of this, or just watching intently? > I mucked around locally, but didn't get anything as far as you have. If you manage to fix some more stuff to make all the tests run, it'd be cool. That's the step 0. Step 1 is to make the patches look nice and polish it, e.g. adding back .new() possibly (but imho using __class__ is a bit cleaner or direct). Step 2 is to prepare patches for review and get it in. I will be busy in the next few weeks, so I am not sure currently when I finish this. Hopefully soon, so that things can get moving. > Cheers > Haz > PS. Do you, or anyone else, know of some "hard" tests for the assumption > system? Something that the old assumption system struggled with would be > nice. I could expand it to a benchmark suite for testing out the new > assumption system capabilities. Well, obviously things like Assume(x+y>5). Ondrej -- You received this message because you are subscribed to the Google Groups "sympy" group. To post to this group, send email to sy...@googlegroups.com. To unsubscribe from this group, send email to sympy+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/sympy?hl=en.
