fix: run hints of each visited pc in run_for_steps

[orizi]

run_for_steps computed the hint slice for the entry pc once and reused it for every step, so hints at subsequent pcs never ran and the entry pc's hints re-ran on every step. Resolve the slice per step, as run_until_pc does. Only the non-extensive_hints path was affected.

Adds a regression test with a vm_enter_scope() hint attached to a non-entry pc.

Based on #2391.

🤖 Generated with Claude Code

---

This change is 

[github-actions]

Benchmark Results for unmodified programs 🚀

Command	Mean [s]	Min [s]	Max [s]	Relative
base big_factorial	2.068 ± 0.033	2.037	2.150	1.00
head big_factorial	2.077 ± 0.060	2.022	2.166	1.00 ± 0.03

Command	Mean [s]	Min [s]	Max [s]	Relative
base big_fibonacci	1.932 ± 0.032	1.907	1.990	1.01 ± 0.02
head big_fibonacci	1.918 ± 0.024	1.902	1.984	1.00

Command	Mean [s]	Min [s]	Max [s]	Relative
base blake2s_integration_benchmark	6.312 ± 0.197	6.192	6.798	1.00
head blake2s_integration_benchmark	6.319 ± 0.121	6.231	6.562	1.00 ± 0.04

Command	Mean [s]	Min [s]	Max [s]	Relative
base compare_arrays_200000	1.979 ± 0.016	1.966	2.015	1.00 ± 0.01
head compare_arrays_200000	1.974 ± 0.007	1.965	1.986	1.00

Command	Mean [s]	Min [s]	Max [s]	Relative
base dict_integration_benchmark	1.382 ± 0.009	1.374	1.401	1.00
head dict_integration_benchmark	1.384 ± 0.010	1.373	1.402	1.00 ± 0.01

Command	Mean [s]	Min [s]	Max [s]	Relative
base field_arithmetic_get_square_benchmark	1.111 ± 0.011	1.097	1.126	1.00
head field_arithmetic_get_square_benchmark	1.127 ± 0.012	1.112	1.153	1.01 ± 0.02

Command	Mean [s]	Min [s]	Max [s]	Relative
base integration_builtins	6.319 ± 0.025	6.293	6.373	1.00
head integration_builtins	6.346 ± 0.016	6.322	6.368	1.00 ± 0.00

Command	Mean [s]	Min [s]	Max [s]	Relative
base keccak_integration_benchmark	6.420 ± 0.032	6.363	6.459	1.00
head keccak_integration_benchmark	6.473 ± 0.072	6.415	6.636	1.01 ± 0.01

Command	Mean [s]	Min [s]	Max [s]	Relative
base linear_search	1.974 ± 0.009	1.962	1.987	1.00
head linear_search	1.992 ± 0.025	1.977	2.062	1.01 ± 0.01

Command	Mean [s]	Min [s]	Max [s]	Relative
base math_cmp_and_pow_integration_benchmark	1.471 ± 0.033	1.449	1.538	1.01 ± 0.03
head math_cmp_and_pow_integration_benchmark	1.463 ± 0.018	1.453	1.514	1.00

Command	Mean [s]	Min [s]	Max [s]	Relative
base math_integration_benchmark	1.437 ± 0.011	1.426	1.459	1.00
head math_integration_benchmark	1.437 ± 0.007	1.428	1.454	1.00 ± 0.01

Command	Mean [s]	Min [s]	Max [s]	Relative
base memory_integration_benchmark	1.191 ± 0.009	1.181	1.205	1.00 ± 0.01
head memory_integration_benchmark	1.188 ± 0.005	1.182	1.195	1.00

Command	Mean [s]	Min [s]	Max [s]	Relative
base operations_with_data_structures_benchmarks	1.517 ± 0.017	1.507	1.565	1.00
head operations_with_data_structures_benchmarks	1.523 ± 0.016	1.511	1.556	1.00 ± 0.02

Command	Mean [ms]	Min [ms]	Max [ms]	Relative
base pedersen	534.9 ± 8.8	525.4	553.6	1.00 ± 0.02
head pedersen	534.5 ± 4.0	525.7	540.9	1.00

Command	Mean [ms]	Min [ms]	Max [ms]	Relative
base poseidon_integration_benchmark	646.4 ± 8.2	641.5	668.6	1.00
head poseidon_integration_benchmark	651.6 ± 21.3	638.0	708.3	1.01 ± 0.04

Command	Mean [s]	Min [s]	Max [s]	Relative
base secp_integration_benchmark	1.828 ± 0.028	1.810	1.903	1.00 ± 0.02
head secp_integration_benchmark	1.824 ± 0.015	1.815	1.865	1.00

Command	Mean [ms]	Min [ms]	Max [ms]	Relative
base set_integration_benchmark	653.5 ± 1.8	650.4	657.2	1.00
head set_integration_benchmark	654.1 ± 4.9	650.7	667.7	1.00 ± 0.01

Command	Mean [s]	Min [s]	Max [s]	Relative
base uint256_integration_benchmark	3.742 ± 0.043	3.716	3.860	1.00
head uint256_integration_benchmark	3.765 ± 0.028	3.734	3.829	1.01 ± 0.01

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 96.23%. Comparing base (c610d41) to head (1ebfbb4).

Additional details and impacted files

@@                           Coverage Diff                           @@
##           claude/vm-execution-performance-32hi0c    #2392   +/-   ##
=======================================================================
  Coverage                                   96.23%   96.23%           
=======================================================================
  Files                                         107      107           
  Lines                                       37985    38009   +24     
=======================================================================
+ Hits                                        36555    36579   +24     
  Misses                                       1430     1430           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[ilyalesokhin-starkware]

-- commits line 7 at r1:
rephrase

Code quote:

  run_for_steps computed the hint slice for the entry pc once and reused
  it for every step, so hints at subsequent pcs never ran and the entry
  pc's hints re-ran on every step. Resolve the slice per step, as
  run_until_pc does. Only the non-extensive_hints path was affected.