[LUR-22] Replace pallas::Scalar as the default test field with BN256

[huitseeker]

Among the instances of CurveCycleEquipped in proof::nova we mostly use the one pegged to pallas::Scalar for all of our tests and as the default test workhorse. Here are the instance counts for this field:

src/symbol.rs:1
src/field.rs:6
src/parser/syntax.rs:1
src/eval/lang.rs:1
src/z_data/z_cont.rs:1
src/public_parameters/mod.rs:1
src/parser/string.rs:1
src/z_data/z_expr.rs:1
src/parser/base.rs:1
src/z_data/z_ptr.rs:3
src/z_data/serde/mod.rs:1
src/coprocessor/trie/mod.rs:1
src/z_data/z_store.rs:1
src/circuit/gadgets/constraints.rs:1
examples/sha256_ivc.rs:1
examples/sha256_nivc.rs:1
src/circuit/circuit_frame.rs:1
examples/circom.rs:1
benches/fibonacci.rs:1
src/lem/store.rs:1
examples/tp_table.rs:1
benches/end2end.rs:1
src/lem/eval.rs:1
benches/public_params.rs:1
benches/sha256.rs:1
src/proof/nova.rs:1
src/proof/tests/nova_tests_lem.rs:1
src/num.rs:2
src/cli/mod.rs:4
src/lem/tests/eval_tests.rs:1
src/lem/tests/misc.rs:1
src/lem/tests/nivc_steps.rs:1

However, we're more likely to use the instance on the BN256 Scalar field as the one that will be most relevant to us:
https://github.com/lurk-lab/lurk-rs/blob/bc136099c58c1402510b949b29768d604642405b/src/proof/nova.rs#L69-L75

We should replace all instances of pallas::Scalar with BN256, (and correspondingly, any instances of pallas::Vesta with grumpkin's Scalar field), which will also help the https://github.com/lurk-lab/solidity-verifier project with test vectors.

Important

We should check this does not impact benchmarks significantly (after adoption of halo2curves 0.5.0, and our integration of grumpkin-msm, we have high hopes that this might indeed be true)

Tip

Resolve #1022 at the same time to ensure non-regression.

Warning

We're picky in how we refer to those scalar fields, don't forget to also update the disallowed-methods in .clippy.toml

This issue is not to be confused with #533 which is about application-level integration of different user-configurable fields. We're here talking about the field used by our unit tests, but we should probably set BN256 as the default nonetheless.

_{From SyncLinear.com | LUR-22}