Some WHATWG streams types not available on `globalThis`

[lucacasonato]

The CI errors seem to be typing issues exposed by this PR:

  ByteLengthQueuingStrategy,
  ~~~~~~~~~~~~~~~~~~~~~~~~~
    at file:///Users/runner/work/deno_std/deno_std/node/stream/web.ts:2:3

TS2339 [ERROR]: Property 'CountQueuingStrategy' does not exist on type 'typeof globalThis'.
  CountQueuingStrategy,
  ~~~~~~~~~~~~~~~~~~~~
    at file:///Users/runner/work/deno_std/deno_std/node/stream/web.ts:3:3

TS2339 [ERROR]: Property 'TextDecoderStream' does not exist on type 'typeof globalThis'.
  TextDecoderStream,
  ~~~~~~~~~~~~~~~~~
    at file:///Users/runner/work/deno_std/deno_std/node/stream/web.ts:8:3

TS2339 [ERROR]: Property 'TextEncoderStream' does not exist on type 'typeof globalThis'.
  TextEncoderStream,
  ~~~~~~~~~~~~~~~~~
    at file:///Users/runner/work/deno_std/deno_std/node/stream/web.ts:9:3

TS2339 [ERROR]: Property 'TransformStreamDefaultController' does not exist on type 'typeof globalThis'.
  TransformStreamDefaultController,
  ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
    at file:///Users/runner/work/deno_std/deno_std/node/stream/web.ts:11:3

TS2339 [ERROR]: Property 'WritableStreamDefaultController' does not exist on type 'typeof globalThis'.
  WritableStreamDefaultController,
  ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
    at file:///Users/runner/work/deno_std/deno_std/node/stream/web.ts:13:3

these are definitely defined on the global so just seem to be omitted from the types for some reason.

Originally posted by @guybedford in denoland/std#1430 (comment)

[kitsonk]

TypeScript only "exposes" vars. Items declared as class or const are not exposed on globalThis/global scope. We would need to declare them as interface WritableStreamDefaultController and declare var WritableStreamDefaultController to have them exposed in the global scope/globalThis.

(ref: microsoft/TypeScript#39504)

[kitsonk]

(though they will appear in the ambient global context, 🤷)

[caspervonb]

I already (partially, only for text streams) patched this a few months ago and we reverted because it was considered breaking iirc? Would be great to finally fix these errors for 2.0 tho.

[bartlomieju]

Is solution from #12429 acceptable?

Ref #5666
Ref #9957

[kitsonk]

Yeah, this is a duplicate of #9957 really, forgot about it.

[kitsonk]

And yes @bartlomieju the approach in #12429 would resolve the issue.