bench: exclude non-deterministic num-primes benchmarks by default

num-primes uses thread_rng() internally, making its benchmarks
non-deterministic and causing false regressions in CodSpeed CI.

Remove num-primes from default features so it's excluded from CI
benchmarks. All num-primes bench_functions are gated behind
#[cfg(feature = "num-primes")] - use --features num-primes to
include them for local comparison.

Author: sylvestre

.github/workflows/codspeed.yml

@@ -31,7 +31,7 @@ jobs:
           bins: cargo-codspeed
 
       - name: Build the benchmark target(s)
-        run: cargo codspeed build -m ${{ matrix.mode }} -p num-prime-bench --no-default-features
+        run: cargo codspeed build -m ${{ matrix.mode }} -p num-prime-bench
 
       - name: Run the benchmarks
         uses: CodSpeedHQ/action@v4

bench/Cargo.toml

@@ -22,4 +22,5 @@ is_prime = "2.0.7"
 glass_pumpkin = "1.2.0"
 
 [features]
-default = ["num-primes"]
+default = []
+num-primes = ["dep:num-primes"]