On Sat, Feb 7, 2009 at 1:58 PM, Florent Hivert <florent.hiv...@univ-rouen.fr> wrote: > > Hi > >> Yes, write code to workaround this case. > > This is mostly done except that on some ring the 0x3 matrix is considered to > be invertible ! And also that the error messages are inconsistent: some raise > a ValueError some raise an ArithmeticError, some says "self must be a square > matrix" some other "matrix has to be square". Since I'm on it right now I can > deal with it. Two questions: > 1 - it's not clear from me which of the two error should be raised.
Regarding inverting a *square* matrix, the error should be ZeroDivisionError. Take a lesson from Python: >>> 1/0 Traceback (most recent call last): ... ZeroDivisionError: integer division or modulo by zero >>> 1.0/0.0 Traceback (most recent call last): ... ZeroDivisionError: float division For trying to invert a non-square matrix, I think the error should be a ValueError, since you're inputing an improper value so that the operation is meaningless. >From http://docs.python.org/library/exceptions.html exception ValueError Raised when a built-in operation or function receives an argument that has the right type but an inappropriate value, and the situation is not described by a more precise exception such as IndexError. exception ArithmeticError The base class for those built-in exceptions that are raised for various arithmetic errors: OverflowError, ZeroDivisionError, FloatingPointError. > 2 - to ensure coherency, it would be easier to have only one file where the > results for different rings and data structures are handled. Do you think > this is a good idea to have a file called says "coherency_tests.py" which > contains only tests and not proper code ? > > My opinion is that this is not the right way to do. I think it fits perfectly > into Nicolas check_... framework. So that in the top class Matrix in > matrix0.py one can make a check_trivial_matrices(self) methods and call check > for all the doc. Unfortunately for this we have to wait for this device to be > integrated. Any though ? I don't really understand what you just wrote. However, factoring out the preconditions for checking whether, e.g., __invert__ should work seems like a good idea to me. This could be either in another file, or just in a cpdef'd method that is defined in matrix0.pyx. self._check_invert_conditions() or something... Using cpdef is a good idea for speed purposes. William --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to sage-devel-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://www.sagemath.org -~----------~----~----~----~------~----~------~--~---
