Improve mock prover errors printing

[zemse]

• fmt Goldilocks(18446744069414584319) as -2 as well as Sum Product expressions.
• include instance id and relevant wit in.

Before

AssertEqualError "test_assert_zero_1/require_equal/a + 1 == b": Left != Right
Left: GoldilocksExt2([Goldilocks(4), Goldilocks(0)])
Right: GoldilocksExt2([Goldilocks(5), Goldilocks(0)])
Left Expression: ScaledSum(WitIn(1), Constant(Goldilocks(1)), Constant(Goldilocks(0)))
Right Expression: Sum(ScaledSum(WitIn(0), Constant(Goldilocks(1)), Constant(Goldilocks(0))), Constant(Goldilocks(1)))

AssertZeroError "test_assert_zero_1/require_zero/c - 2 == 0": Evaluated expression is not zero
Expression: Sum(ScaledSum(WitIn(2), Constant(Goldilocks(1)), Constant(Goldilocks(0))), Constant(Goldilocks(18446744069414584319)))
Evaluation: GoldilocksExt2([Goldilocks(18446744069414584322), Goldilocks(0)])

LookupError "test_lookup_1/assert_u5/assert u5": Evaluated expression does not exist in T vector
Expression: ScaledSum(WitIn(0), Challenge(1, 1, GoldilocksExt2([Goldilocks(1), Goldilocks(0)]), GoldilocksExt2([Goldilocks(0), Goldilocks(0)])), Challenge(0, 1, GoldilocksExt2([Goldilocks(1), Goldilocks(0)]), GoldilocksExt2([Goldilocks(0), Goldilocks(0)])))
Evaluation: GoldilocksExt2([Goldilocks(1000001), Goldilocks(0)])

After

AssertEqualError "test_assert_zero_1/require_equal/a + 1 == b"
Left: GoldilocksExt2[4,0] != Right: GoldilocksExt2[5,0]
Left Expression: 1 * WitIn(1) + 0
Right Expression: 1 * WitIn(0) + 0 + 1
Inst[0]: WitIn(1)=4,WitIn(0)=4

AssertZeroError "test_assert_zero_1/require_zero/c - 2 == 0": Evaluated expression is not zero
Expression: 1 * WitIn(2) + 0 + (-2)
Evaluation: GoldilocksExt2[1,0] != 0
Inst[1]: WitIn(2)=3

LookupError "test_lookup_1/assert_u5/assert u5": Evaluated expression does not exist in T vector
Expression: Challenge(1) * WitIn(0) + Challenge(0)
Evaluation: GoldilocksExt2[1000001,0]
Inst[1]: WitIn(0)=1000

[hero78119]

Awesome job for improving the readiblity of error message 👍 👍

Just left some optionally optimisation suggestion

• some imperative programming style e.g. alloc vec![]; for xxxx {push sth to vec} can either Vec::with_capacity() in advance or use functional style iter().map(||xxxxx).collect()
• some left over comments can be clean up

[hero78119]

LGTM 👍