fix: optimize zero constant detection and assignment

[ivokub]

Description

Found in fuzzing test:

type C1_6314_8435483601975321392 struct {
	X0V frontend.Variable
	Y0V frontend.Variable `gnark:",public"`
}

func (circuit *C1_6314_8435483601975321392) Define(api frontend.API) error {
	x1 := api.IsZero(circuit.X0V) // depends on X0V
	x2 := api.And(0, x1)          // always 0
	x3 := api.Cmp(x2, 1)          // always -1
	api.AssertIsEqual(1, x3)      // always false
	return nil
}

type C2_6314_8435483601975321392 struct {
	X0V frontend.Variable
	Y0V frontend.Variable `gnark:",public"`
}

func (circuit *C2_6314_8435483601975321392) Define(api frontend.API) error {
	x1 := api.Cmp(circuit.X0V, circuit.X0V) // always 0
	x2 := api.Cmp(x1, -1)                   // always -1
	x3 := api.IsZero(x2)                    // always 0
	x4 := api.IsZero(circuit.X0V)           // depends on X0V
	x5 := api.IsZero(circuit.X0V)           // depends on X0v
	x6 := api.Select(x3, x4, x5)            // always returns x5
	x7 := api.And(0, x6)                    // always 0
	api.AssertIsLessOrEqual(1, x7)          // always fails
	return nil
}

Where previously R1CS compiler didn't fail for second circuit but did for SCS.

Type of change

• Bug fix (non-breaking change which fixes an issue)

How has this been benchmarked?

Ran BW6 in BN254 Groth16 verifier circuit to benchmark compile time (as we now do big loop for constant value checking).

              │  old.txt   │           new.txt           │
              │   sec/op   │   sec/op    vs base         │
BW6InBN254-16   9.434 ± 2%   9.501 ± 1%  ~ (p=0.589 n=6)

              │   old.txt    │            new.txt            │
              │     B/op     │     B/op      vs base         │
BW6InBN254-16   21.28Gi ± 0%   21.28Gi ± 0%  ~ (p=1.000 n=6)

              │   old.txt   │           new.txt            │
              │  allocs/op  │  allocs/op   vs base         │
BW6InBN254-16   91.48M ± 0%   91.48M ± 0%  ~ (p=0.937 n=6)

Checklist:

• I have performed a self-review of my code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• I have added tests that prove my fix is effective or that my feature works
• I did not modify files generated from templates
• golangci-lint does not output errors locally
• New and existing unit tests pass locally with my changes
• Any dependent changes have been merged and published in downstream modules