~30% faster, ~48% less memory used, and ~37% fewer allocations on our simple benchmark. **IMPORTANT**: **this does _not_ mean that `oxlint` type-aware linting will run this much faster.** When I ran this with `oxlint`, it was not noticeably faster, it was maybe 1% better. But it was consistently faster, so I think it's worth merging. The intuition for the small real-world speedup is that the bottleneck is in the program creation and reading files from the filesystem, not the linting part itself.

The main idea is that we are currently re-allocating the rule context for every single file and every single rule. In most cases, though the behavior is exactly the same, we are just changing what file are rule we are linting with.

I refactored the diagnostic reporting code to be such that it never reallocates the diagnostic reporting functions: instead, we update the rule name and file and mutate the diagnostic reported with the current state after we've received it. This means we don't need to allocate this struct over and over again, which should reduce pressure on the GC quite a bit (this does show up in the benchmark too).

Before:

```
goos: darwin
goarch: arm64
pkg: github.com/typescript-eslint/tsgolint/cmd/tsgolint
cpu: Apple M1
BenchmarkAllRulesHeadless-8               	    4495	   2604683 ns/op	 7640387 B/op	   83526 allocs/op
BenchmarkAllRulesHeadless-8               	    4442	   2599917 ns/op	 7675330 B/op	   83833 allocs/op
BenchmarkAllRulesHeadless-8               	    4574	   2590637 ns/op	 7669025 B/op	   83771 allocs/op
BenchmarkAllRulesHeadless-8               	    4442	   2686293 ns/op	 7647700 B/op	   83587 allocs/op
BenchmarkAllRulesHeadless-8               	    4291	   2728212 ns/op	 7654198 B/op	   83647 allocs/op
BenchmarkAllRulesHeadlessSingleThread-8   	    2041	   5863713 ns/op	 7629003 B/op	   83638 allocs/op
BenchmarkAllRulesHeadlessSingleThread-8   	    2029	   5873142 ns/op	 7622800 B/op	   83581 allocs/op
BenchmarkAllRulesHeadlessSingleThread-8   	    2034	   5863612 ns/op	 7614713 B/op	   83508 allocs/op
BenchmarkAllRulesHeadlessSingleThread-8   	    2018	   5881180 ns/op	 7630698 B/op	   83649 allocs/op
BenchmarkAllRulesHeadlessSingleThread-8   	    2038	   5890125 ns/op	 7635571 B/op	   83689 allocs/op
```

After:

```
goos: darwin
goarch: arm64
pkg: github.com/typescript-eslint/tsgolint/cmd/tsgolint
cpu: Apple M1
BenchmarkAllRulesHeadless-8               	    6270	   1827035 ns/op	 4372362 B/op	   52719 allocs/op
BenchmarkAllRulesHeadless-8               	    6024	   1794715 ns/op	 4372635 B/op	   52719 allocs/op
BenchmarkAllRulesHeadless-8               	    6276	   1795347 ns/op	 4354735 B/op	   52559 allocs/op
BenchmarkAllRulesHeadless-8               	    6327	   1837421 ns/op	 4389465 B/op	   52866 allocs/op
BenchmarkAllRulesHeadless-8               	    6244	   1863864 ns/op	 4375612 B/op	   52741 allocs/op
BenchmarkAllRulesHeadlessSingleThread-8   	    2756	   4325317 ns/op	 3948056 B/op	   52395 allocs/op
BenchmarkAllRulesHeadlessSingleThread-8   	    2791	   4327110 ns/op	 3953848 B/op	   52463 allocs/op
BenchmarkAllRulesHeadlessSingleThread-8   	    2770	   4336311 ns/op	 3924766 B/op	   52215 allocs/op
BenchmarkAllRulesHeadlessSingleThread-8   	    2758	   4342168 ns/op	 3948428 B/op	   52401 allocs/op
BenchmarkAllRulesHeadlessSingleThread-8   	    2726	   4339023 ns/op	 3954526 B/op	   52460 allocs/op
```