[WIP] Deprecate solve_linear_system for linsolve

[aktech]

TODO

• Return dictionary of linsolve solution
• [WIP] Deprecate solve_linear_system for linsolve #10000 (comment)
• Fix Failing Tests

@asmeurer @hargup

[asmeurer]

This should be 0.7.7 (actually 1.0, but we'll have to change all instances of the former to the latter at some point anyway).

[asmeurer]

Nice, issue 10000!

[asmeurer]

The tests need to be fixed.

[asmeurer]

It's good to see this by the way. I think this fixes #4949.

[aktech]

I think this fixes #4949.

There are few more functions which duplicates solving linear equations, will try to finish this soon.

Here are all the permutations:

solve(equations, unknowns)
solve_linear_system(aug_matrix)
solve_linear_system_LU(aug_matrix, unknowns)
solve_poly_system(equations, unknowns)
solve_linear(single_equation)
polys.solvers.solve_lin_sys(equations, ring)
minsolve_linear_system(system, symbols)

[asmeurer]

Is the call signature different for linsolve (this function uses *symbols)? If so, you might mention that in the deprecation notice.

[aktech]

Indeed.

[aktech]

The **flags have been removed.

[aktech]

It seems there are some critical issues with rref which effects linsolve. If we have to release 1.0 soon, then we should skip deprecating this for now. We should just document that solve_linear_system will be replaced by linsolve and people should use linsolve. Similiar to what we would do with solveset (linsolve is a part of solveset module).