Christopher Smith

restarted tests

I think it is the sphynx test that is holding this up. Maybe try rebase with current master?

I don't know which test is holding this up. And the rebase looks fine.

```python >>> simplify(S.Half*log(4)) log(2) ``` `logombine` is responsible for making this change. Here is a less fortuitous change: ```python >>> logcombine(65*ln(49/S(50))/24) log(16807*2**(7/24)*5**(7/12)*7**(5/12)/125000) ``` In simplify, `logcombine` is attempted without any...

I wonder if this would be a good candidate for the newly added routine to find items that are connected in a tree.

Has Aaron's concern from issue page been addressed: > We should be careful that the modulus is always real (or at least reasonably canonical if it's symbolic). Otherwise the output...

Otherwise this looks good.

> These expressions involving only a single radical and I are pretty reasonable and in some situations probably easier to work with or more natural than complex exponentials. For multiples...

The new distance method does not resolve this problem: ```python plane = Plane((0,0,0),(1,0,0),(1,1,0)) l1=Line((x,y,z),(1,1,1)) l3=Line((x,x,y),(1,1,1)) # find image of lines in plane l1p=Line(*[plane.projection(i) for i in l1.args]) # Line3D(Point3D(x, y,...

It would be a good idea to simply raise a NotImplementedError for symbolic cases that cannot be solve.