On Mar 5, 12:04 pm, smichr <smi...@gmail.com> wrote: > > This is False for reasons I am trying to track down. The problem > > doesn't exist in master > > Correction: the problem exists in master. (The original raising- > expression wasn't causing a problem in master but was in 1766, but the > underlying problem which has been distilled to the issue shown in the > previous comment is something that exists in master.) I'll open an > issue for it.
OK, that's been fixed. I wrote a tool to see if lines that have been added are covered and wrote tests for lines that weren't covered...but that was a day or two ago. Everything that is ready for review is before the "quartz ----------" commit. (Quartz is in pretty good shape, but there are still some issues to flesh out.) I've tried to put related patches in close proximity. There were some questions about some of the things that I've done, e.g. igcd; making Rational work on basic expressions (Rational(.3*x+7)- >3*x/10+7; changing how integral atoms are reported. I also wrote a hint manager and used that with expand() but a concern was raised that this slowed things down. But since then I have cleaned up expand a lot so redundant calls are reduced. My run time for the suite is generally less than 10 minutes and it was creeping up toward 20 minutes. I would be very interested to see if this is a real effect or if I just happened to always be doing something else that slowed the computer down while running the tests. With the non-commutative coeff and the issue with the sign (1849?) fixed, I don't anticipate changing anything in the pre-quartz commits now. So it would be nice to get the review process finished on those. All tests for the whole branch pass; it is possible that with the re- ordering of patches that intermediate commits will have failing tests. It would be nice if these could be checked with a faster computer as there are about about 100 commits starting with 1789: extra comments in mr_safe, not including the quartz-related commits. All doctests do not pass: I don't have numpy installed and some matrix doctests require that; there is a failure involving expand that I have to look into; with the improved solver, lots of ode's are coming back with solutions for f(x)...it would be nice to have a "leave_implicit" hint so f(x) wouldn't be solved for explicitly otherwise some of the solutions are quite long. I'll post when I've fixed the expand issue. -- You received this message because you are subscribed to the Google Groups "sympy-patches" group. To post to this group, send email to sympy-patc...@googlegroups.com. To unsubscribe from this group, send email to sympy-patches+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/sympy-patches?hl=en.