Oscar Benjamin

> We really probably should just make Symbol more subclass friendly for these kinds of use-cases since people are doing them all the time anyways. It can be more subclass...

It is safe to cache the assumptions between different instances of Symbol according to the way that Symbol itself is designed which is that its assumptions are all fixed at...

Have you considered that you can make a `MySymbol` class by subclassing `Expr` rather than `Symbol`? The assumptions system is really not supposed to be used for the case that...

You can do it like this: ```python from __future__ import annotations from sympy import Expr, sin from typing import Hashable class MySymbol(Expr): name: str data: Hashable # This makes differentiation...

@mloubout where did this end up? I'm contemplating a 1.14.1 bugfix release after gh-28157.

There won't be any change to this in that release unless some change is agreed here and then backported.

We have this which makes sense: ```python In [1]: from sympy.polys.polytools import factor_system In [2]: a, b, x, y = symbols('a, b, x, y') In [3]: eqs = [(a -...

The function here does not quite work correctly: ```python In [17]: factor_system([(x - 1)*(y - 2), (y - 2)*(z - 3)]) Out[17]: [{x - 1, z - 3}, {x -...

> NotImplementedError: multivariate polynomials over finite fields That's fine. The underlying polys module does not support factorising multivariate polynomials over finite fields as the error message says. This feature is...

Test cases should be added for things like `(sin(1)**2 + cos(1)**2 - 1)` and `cos(x)**2 - sin(x)**2`.