Hi. This is great. I thought you had abandoned the work. It's good to see that you are picking it up again.
On Sun, Nov 27, 2011 at 6:52 PM, Mateusz Paprocki <[email protected]> wrote: > Hi, > > On 28 November 2011 02:01, Kate MacInnis <[email protected]> wrote: >> >> So last spring, I started working on removing the auto-distribution of >> constants. Life got busy, and I had to set that project aside for >> awhile. When I decided to come back to it over the long weekend, I >> decided to start fresh. I've now got almost all tests passing. :-) > > Thanks for working on this. > >> >> But I do have a couple of questions. >> >> Awhile back, there was talk of a hint manager for expand. Whatever >> happened with that? Was it that the behavior couldn't be agreed on, >> or was it just a matter of getting busy with other things? I've >> implemented a simple one as part of my solution. If any hint is set >> to True, it runs only that hint. > > Being busy is one thing, but the real problem is that expand()'s API is just > to complex and adding a hint manager may not solve our problems. Also, a > sort of option manager was implemented in sympy.polys.polyoptions and can be > adapted for usage in expand(). I think the hint manager that Chris was working on ended up being bogged down with other code, and no one ever got to reviewing it. The simple thing you explain should be a good enough start. If you want to adapt the Polys hint manager, that would be nice. But as Mateusz said, we need to refactor all of expand() anyway. > >> >> Also, on the two tests that I haven't gotten to pass yet: >> >> I'm getting an AttributeError exception in integrals/tests/ >> test_risch.py, but if I run the test from inside isympy it doesn't >> have a problem. Any ideas? > > I don't get this failure when using your branch. Can you post output from > bin/test (complete with architecture, seed, ground types, etc.)? Do you test > the individual file or the whole test suite (are there any differences in > outcome)? Also, see if it makes a difference if the cache is enabled or not. It is enabled by default. You can disable it by passing -C to isympy or to ./bin/test. Note that there are test failures in master with the cache off, so if you find a failure with that, see if it's in master too. Also note that things are a lot slower with the cache off (especially in expand() code), so I recommend not trying to run the whole test suite with the cache off. Aaron Meurer > >> >> The only other test that I'm having a problem with is test_formulae in >> simplify/tests/test_hyperexpand.py. It has a loop where it >> numerically tests each formula in FormulaCollection. Are there any >> resources anyone can point me to as I wrap my head around this one? > > I reproduced the problem. I will look into this. > >> >> If anyone wants to take a look at it, here's my working branch: >> https://github.com/kmacinnis/sympy/tree/remove-auto-distribution I >> apologize in advance if the commit history is a little screwy-- when I >> tried to rebase on the most recent updates things got a little mangled >> up. >> >> Thanks in advance for any feedback. >> >> -- >> You received this message because you are subscribed to the Google Groups >> "sympy" group. >> To post to this group, send email to [email protected]. >> To unsubscribe from this group, send email to >> [email protected]. >> For more options, visit this group at >> http://groups.google.com/group/sympy?hl=en. >> > > Mateusz > > -- > You received this message because you are subscribed to the Google Groups > "sympy" group. > To post to this group, send email to [email protected]. > To unsubscribe from this group, send email to > [email protected]. > For more options, visit this group at > http://groups.google.com/group/sympy?hl=en. > -- You received this message because you are subscribed to the Google Groups "sympy" group. To post to this group, send email to [email protected]. To unsubscribe from this group, send email to [email protected]. For more options, visit this group at http://groups.google.com/group/sympy?hl=en.
