chore: use dyn DatabaseExt everywhere

[DaniPopes]

Closes #8404

[klkvr]

wondering if we could get a noticable perf degradation after this? given that Inspector::step will be dynamically dispatched now?

[DaniPopes]

Looks like about 10% slower

---

Solady:

Benchmark 1: forge-main test
  Time (mean ± σ):      1.160 s ±  0.107 s    [User: 23.772 s, System: 0.174 s]
  Range (min … max):    1.012 s …  1.330 s    10 runs

Benchmark 2: forge-dyn test
  Time (mean ± σ):      1.246 s ±  0.094 s    [User: 25.132 s, System: 0.212 s]
  Range (min … max):    1.104 s …  1.396 s    10 runs

Summary
  forge-main test ran
    1.07 ± 0.13 times faster than forge-dyn test

Sablier v2 core:

Benchmark 1: forge-main test --nmc Fork
  Time (mean ± σ):      8.526 s ±  0.394 s    [User: 49.358 s, System: 0.754 s]
  Range (min … max):    7.992 s …  9.160 s    10 runs

Benchmark 2: forge-dyn test --nmc Fork
  Time (mean ± σ):      9.288 s ±  0.848 s    [User: 55.514 s, System: 0.697 s]
  Range (min … max):    8.306 s … 10.749 s    10 runs

Summary
  forge-main test --nmc Fork ran
    1.09 ± 0.11 times faster than forge-dyn test --nmc Fork

[klkvr]

I think hottest paths are using &mut InspectorStack, so we could cheaply improve this by using I: InspectorExt?

at the cost of bloat from InspectorStack and dyn InspectorExt (from CheatcodesExecutor)

[DaniPopes]

Yeah I'd be fine with that

[klkvr]

implemented in #9011