fix(fuzz): Handle overflow errors from the elaborator

[aakoshh]

Description

Problem*

Resolves the error handling causing nightly failure in https://github.com/noir-lang/noir/actions/runs/15866606791/job/44734570793

Summary*

The comptime_vs_brillig_direct target did not inspect the error returned by the Elaborator to check if it was potentially an arithmetic overflow. Doing so allows us to go ahead and compare it with the runtime result, and not fail a test.

Additional Context

This was the test reproduction:

❯ NOIR_AST_FUZZER_SHOW_AST=1 NOIR_AST_FUZZER_SEED=0x863f31e800001400 cargo test -q -p noir_ast_fuzzer_fuzz comptime_vs_brillig_direct
   
running 1 test
test targets::comptime_vs_brillig_direct::tests::fuzz_with_arbtest ... FAILED

failures:

---- targets::comptime_vs_brillig_direct::tests::fuzz_with_arbtest stdout ----
---
unconstrained fn main() -> pub i8 {
    let mut ctx_limit: u32 = 5;
    func_1_proxy(("EFFW", -136065753279342490752102008551958553268, "DZBK", 113, "NSBG"), ctx_limit)
}
unconstrained fn func_1(mut a: (str<4>, Field, str<4>, i8, str<4>), _ctx_limit: &mut u32) -> i8 {
    (-if true {
        if false {
            a.3
        } else {
            -128
        }
    } else {
        -40
    })
}
unconstrained fn func_1_proxy(mut a: (str<4>, Field, str<4>, i8, str<4>), mut ctx_limit: u32) -> i8 {
    func_1(a, (&mut ctx_limit))
}

---

thread 'targets::comptime_vs_brillig_direct::tests::fuzz_with_arbtest' panicked at /Users/aakoshh/Work/aztec/noir/tooling/ast_fuzzer/src/compare/comptime.rs:92:17:
elaborator error while interpreting generated comptime code: Compile([InterpreterError(NegateWithOverflow { location: Location { span: Span(Span { start: ByteIndex(322), end: ByteIndex(448) }), file: FileId(0) } })])
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

arbtest failed!
    Seed: 0x863f31e800001400

Documentation*

Check one:

• No documentation needed.
• Documentation included in this PR.
• [For Experimental Features] Documentation to be submitted in a separate PR.

PR Checklist*

• I have tested the changes locally.
• I have formatted the changes with Prettier and/or cargo fmt on default settings.

[github-actions]

⚠️ Performance Alert ⚠️

Possible performance regression was detected for benchmark 'Execution Time'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.20.

Benchmark suite	Current: 9297c78	Previous: 9fc413d	Ratio
rollup-merge	0.004 s	0.003 s	1.33

This comment was automatically generated by workflow using github-action-benchmark.

CC: @TomAFrench

[rkarabut]

So this one wasn't prevented by avoid_overflow then huh 🤔

[aakoshh]

Yeah I don't think gen_unary looks at avoid_overflow, that's a good point.