chore: Change AST fuzzer recursion limit

[aakoshh]

Description

Problem*

Resolves intermittent slow tests in #8048

Summary*

Replaces mut ctx_depth: u32 with ctx_limit: &mut u32, to handle exponential cases similar to Fibonacci by limiting the overall number of recursive calls before shortcutting functions to return a random value, rather than allowing the maximum depth to be reached.

The complication is that it's not possible to pass a mutable reference from ACIR to Brillig. To overcome this we: create a func_{i}_proxy for unconstrained functions, which takes mut ctx_limit: u32 by value and passes it to func_{i} by reference.

Additional Context

Here's an example of a function which was slow with ctx_depth, rewritten to use mutable references:

Slow due to recursion

global G_A: [(Field, Field, bool, bool); 0] = [];
global G_B: [[u16; 0]; 3] = [[], [], []];
unconstrained fn main() -> pub (Field, Field, bool, bool) {
    let mut ctx_limit = 5;
    (210599614764714653382489015013312909866, (func_2(func_5("", &mut ctx_limit), if false {
        (!true)
    } else {
        func_5("", &mut ctx_limit)
    }) as Field), func_5(if true {
        ""
    } else {
        ""
    }, &mut ctx_limit), if func_5("", &mut ctx_limit) {
        (!func_5("", &mut ctx_limit))
    } else {
        if func_5(if (!false) {
            ""
        } else {
            func_4([("", false, 112173205673011516923171613784418000313, "ABC", -8147)], &mut ctx_limit).0
        }, &mut ctx_limit) {
            false
        } else {
            func_5("", &mut ctx_limit)
        }
    })
}
fn func_1(mut a: [[u16; 0]; 3], b: (Field, Field, bool, bool), mut c: Field) -> bool {
    b.2
}
unconstrained fn func_2(a: bool, mut b: bool) -> u16 {
    12744
}
fn func_3(a: bool, mut b: [str<0>; 3], mut c: [(u64, bool, u16, u32, bool); 2], ctx_limit: &mut u32) -> Field {
    if (*ctx_limit == 0) {
        -91272772969658566208753441772487222731
    } else {
        *ctx_limit = *ctx_limit - 1;
        if unsafe { func_5_proxy(if unsafe { func_5_proxy(b[0], *ctx_limit) } {
            unsafe { func_4_proxy([("", c[0].1, (a as Field), "IEF", 15837)], *ctx_limit) }.0
        } else {
            ""
        }, *ctx_limit) } {
            61819793148384237204158051638397851904
        } else {
            for idx_d in 8844 .. 8849 {
                if (if unsafe { func_5_proxy("", *ctx_limit) } {
                    51123439257234248396303652390047952106
                } else {
                    (unsafe { func_4_proxy([("", true, -332984236243826044317475693288591622591, "LGQ", 7313)], *ctx_limit) }.1.1 as u128)
                } <= (unsafe { func_2(false, c[0].4) } as u128)) {
                    c[0] = {
                        b = b;
                        let e = "S";
                        (12257872272973752653, true, 29144, (((c[1].2 % 9265) << 125) as u32), a)
                    };
                } else {
                    ()
                };
                ()
            };
            -214491953683494567084585875246146393402
        }
    }
}
unconstrained fn func_4_proxy(mut a: [(str<0>, bool, Field, str<3>, i16); 1], mut ctx_limit: u32) -> (str<0>, (i64, Field), [u16; 0]) {
    func_4(a, &mut ctx_limit)
}
unconstrained fn func_4(mut a: [(str<0>, bool, Field, str<3>, i16); 1], ctx_limit: &mut u32) -> (str<0>, (i64, Field), [u16; 0]) {
    if (*ctx_limit == 0) {
        ("", (-6198611486336017755, -158532646483603675733802980870644575904), [])
    } else {
        *ctx_limit = *ctx_limit - 1;
        (func_4(a, ctx_limit).0, (((a[0].4 >> if true {
            234
        } else {
            34
        }) as i64), (a[0].1 as Field)), if true {
            func_4(a, ctx_limit).2
        } else {
            G_B[0]
        })
    }
}
unconstrained fn func_5_proxy(mut a: str<0>, mut ctx_limit: u32) -> bool {
    func_5(a, &mut ctx_limit)
}
unconstrained fn func_5(mut a: str<0>, ctx_limit: &mut u32) -> bool {
    if (*ctx_limit == 0) {
        true
    } else {
        *ctx_limit = *ctx_limit - 1;
        if func_5(func_4([("", true, -68801463734466096185118923994098536602, "LJH", -16314)], ctx_limit).0, ctx_limit) {
            a = if func_5("", ctx_limit) {
                if func_5(a, ctx_limit) {
                    a = a;
                    {
                        let mut idx_b = 0;
                        loop {
                            if (idx_b == 10) {
                                break
                            } else {
                                idx_b = (idx_b + 1);
                                {
                                    let mut idx_c = 0;
                                    while func_5(a, ctx_limit) {
                                        if (idx_c == 10) {
                                            break
                                        } else {
                                            idx_c = (idx_c + 1);
                                            a = "";
                                        }
                                    }
                                };
                            }
                        }
                    };
                    a = a;
                };
                ""
            } else {
                a
            };
            true
        } else {
            false
        }
    }
}

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 'Test Suite Duration'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.20.

Benchmark suite	Current: 8e8b07b	Previous: 17958e3	Ratio
test_report_noir-lang_noir_bigcurve_	297 s	246 s	1.21

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

CC: @TomAFrench

[rkarabut]

Can we add a test to check if the proxy harness looks like it's supposed to in the resulting AST?

[aakoshh]

@rkarabut done, thanks for flagging that up, I'm sure it's going to be helpful in the future 👍

[github-actions]

⚠️ Performance Alert ⚠️

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

Benchmark suite	Current: 7b02b62	Previous: 76cc8b4	Ratio
rollup-block-root	1710 MB	1420 MB	1.20

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

CC: @TomAFrench

[github-actions]

⚠️ Performance Alert ⚠️

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

Benchmark suite	Current: df0121b	Previous: 6c4dd0c	Ratio
private-kernel-inner	2.754 s	2.204 s	1.25

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

CC: @TomAFrench