Jason K. Moore
Jason K. Moore
Maybe one more question. If we are releasing a bugfix 1.12.1 then should it be setup to work with mpmath >=1.1.0 without an upper bound? Having that upper bound is...
> I think maybe the idea is there might be more breaking changes by the final release? This hits on one of the main issues with upper bounds: we shouldn't...
I just played around with `MatrixSolve()` and it doesn't quite look like what we really need (yet). Some notes: - We should be able to use the MatrixSolve and its...
Another issue is that this is only useful if you lambdify the result and use the numpy printer.
You haven't explained what a block matrix is and why it is needed. I'm guessing this is some way to avoid the need to index `MatrixSolve()`. Is that right?
There are some tests that do seem to work with printing for a bare matrixsolve, e.g. : https://github.com/sympy/sympy/blob/84bc583d9427b781f1a5dcef783767b6b69aa510/sympy/printing/tests/test_octave.py#L265, so maybe the intention is that it is used as an expression.
Here is the PR that introduced MatrixSolve: https://github.com/sympy/sympy/pull/17041 it is more or less a proof of concepts and works in some narrow cases.
I thought about this more last night. If we hard code the use of MatrixSolve here in KanesMethod, then we are stuck with it. So we should be darn sure...
Given that we ran out of time to implement a more general approach to having delayed linear solves, this may be a good idea to merge. I think we should...
Hi @tjstienstra I"m still interested in this PR and would like to keep it open. Is there a reason you deleted it?