Christopher Smith

I wonder if some of the ode related issues will stem from the assumption that this partial extraction is not going to work. So the code would have to be...

I would use something like this... ```diff combines = b.is_positive and e.is_real or b.is_nonnegative and e.is_nonnegative - + if not combines: + # check for partial extraction as in (x**3).subs(x**2,...

perhaps check here: https://github.com/sympy/sympy/blob/258278558b3946e1b10f9aae12b58301da8e5c99/sympy/polys/polytools.py#L6017 Maybe that should be ``` if arg.is_Pow or arg.func is exp: base, exp = arg.as_base_exp() ```

blocked by #23891

Does anyone have an opinion about whether simplification of the undetermined coefficients should be retained? As a return value from poly solvers, it is no longer simplified and that is...

> I think it's better not to simplify. OK, then I have nothing more planned for this PR.

Thanks for the help. We're making progress.

> what should be the fix for the test see #23918

The `real_roots` routine should be used for polynomial, I believe: ```python >>> [i.n(3) for i in real_roots(poly.diff()) if i>0 and i < 1] [0.0911, 0.402, 0.649] ```

Is there a better way to do this? If not, perhaps ```python def sdiv(p, q): """return w, r if p = w*q + r else 0, p Examples ======== >>>...