SquinU3ToClifford on scf.IfElse

[johnzl-777]

I just realized from a comment @david-pl left on #510 that SquinU3ToClifford currently does not handle cases like the following:

# If the measurement result is True,
# apply Rz to a single qubit
if meas_res:
	squin.rz(pi, q[0])

This could easily be interpreted as a CZ.

One could just have some additional logic in the IfToStim rule to handle this but I think the more architecturally correct way to go about this is to bump SquinU3ToClifford up on the chain of rewrites in the SquinToStimPass and THEN let the other rules like IfToStim handle what's left over.

[weinbe58]

This should be a separate simplification. we shouldn't put everything in one rewrite rule.

[johnzl-777]

Then the second method I mentioned should be better, right? We let SquinU3ToClifford have some additional logic to go into loops so we can reuse as much of that logic and then let the IfToStim rewrite rule operate on any cases where you have an scf.IfElse enclosing a pauli gate