refactor(http): prepare for noUncheckedIndexedAccess

[gabelluardo]

Ref: #4040

[iuioiua]

Hi Gabriele, thank you for the PR. PTAL at #4275 (review). It details how these PRs for #4040 should be approached.

[gabelluardo]

Hi @iuioiua, thanks for the suggestion, I think I did it better now.

However, I have a doubt about these two functions:

https://github.com/denoland/deno_std/blob/083a2b12ba45381c78dbabfcf06d48e2da27f16b/http/_negotiation/media_type.ts#L54-L69

and

https://github.com/denoland/deno_std/blob/083a2b12ba45381c78dbabfcf06d48e2da27f16b/http/_negotiation/media_type.ts#L71-L86

I tested with unit tests and seems theses two function can be replaced with accept.split(","); and str.split(";").map((p) => p.trim()). Is a test bug or we can remove the for loops?

[iuioiua]

Hi @iuioiua, thanks for the suggestion, I think I did it better now.

However, I have a doubt about these two functions:

https://github.com/denoland/deno_std/blob/083a2b12ba45381c78dbabfcf06d48e2da27f16b/http/_negotiation/media_type.ts#L54-L69

and

https://github.com/denoland/deno_std/blob/083a2b12ba45381c78dbabfcf06d48e2da27f16b/http/_negotiation/media_type.ts#L71-L86

I tested with unit tests and seems theses two function can be replaced with accept.split(","); and str.split(";").map((p) => p.trim()). Is a test bug or we can remove the for loops?

Yeah, you're mostly right. Each of these can be replaced by header.split(delimiter).map((p) => p.trim()). I don't see how quote counting is needed here.

[iuioiua]

LGTM. @kitsonk, please feel free to review as this code was introduced to the codebase by you.