Pin the hash seed in the execution benchmark harness - #9235

Split out of #9224 at @fanninpm's request.

Change

bench_rustpython_code in benches/execution.rs now sets settings.hash_seed = Some(0). benches/microbenchmarks.rs has used the same setting since #9126.

Every RustPython interpreter in this harness is created by that function. The hash secret is set once per process, by the first interpreter, so all execution and pystone benchmarks now hash str/bytes the same way on every run. Dict and set probe sequences no longer change between runs.

Notes

• Only benchmark code changes. Interpreter behaviour is the same.
• cargo check --bench execution passes.
• This does not stabilise parse_to_ast::*. Those benchmarks only run the ruff parser and never hash Python objects. Their swings come from the heap state that earlier benchmarks leave behind, and that depends on the HashMap iteration order in criterion_benchmark.

Comment @codspeedbot make the execution benchmarks run in a deterministic order and I'll add that to this PR.