experimental: UseRequest/ResponseBody methods

[M4tteoP]

Exposing as experimental two new methods to provide the bodies to Coraza. They are meant to be used when the entire request body is available, as it avoids the need for an extra copy into the body buffers.

Benchmark:

goos: darwin
goarch: arm64
pkg: github.com/corazawaf/coraza/v3/internal/corazawaf
BenchmarkUseRequestBody/WriteRequestBody-10         	    1383	    912582 ns/op	10544513 B/op	     214 allocs/op
BenchmarkUseRequestBody/UseRequestBody-10           	   63420	     17178 ns/op	   50075 B/op	     212 allocs/op
PASS
ok  	github.com/corazawaf/coraza/v3/internal/corazawaf	3.559s

Usage:

tx := waf.NewTransaction().(corazaexperimental.Transaction)
body := []byte("body") // Retrieved from the connector or somewhere
tx.UseResponseBody(body)

Usage:

• Go-ftw within its new quantitative testing phase: coreruleset/go-ftw@main...use_body
• Experimenting with new connectors (no code available, yet)

[jptosso]

My only concern is maintaining a pointer that can be flushed at any point.
I'm not sure if telling the user to avoid using b slice is enough

[M4tteoP]

My only concern is maintaining a pointer that can be flushed at any point.
I'm not sure if telling the user to avoid using b slice is enough

Hey, I elaborated on the description a bit more. However, this feature fundamentally relies on this approach, so I would assume that anyone using it is intentionally willing to let Coraza the ownership of that pointer

[jcchavezs]

I am missing the use cases where this can be used. In regular go code I think it is risky as mutations can happen, in other environments I think is more acceptable. I would not merge this until we show solid code using it.