Request headers support for protocols

[Gozala]

This is followup from #2 where you can find all the details. Below quote is short summary:

Request Headers - Things seems here sadly far more complicated, but it is also clear that this is really important for proper support of Video / Audio tags that just assume HTTP backend. It was suggested to me to that we could do following:

• Plain data from request so that protocol handler can parse HTTP headers or alternative encoding of metadata.
  - Expose existing HTTP parser so it could be used. There was quite a reluctance on this though for reasons I did not understand (Part of it was that it's written in C++ and could not be easily extracted for just parsing use). It was highly recommended to instead just support more strictly encoded subset of the headers as anything send from video / audio tags would fit.

Unfortunately it seems that supporting request headers would require C++ work and landing corresponding changes into Firefox which is to say it's likely going to take a while and likely will come after we illustrated that people are actually building on this work.

[lidel]

Is analog to HTTP's Range requests tracked under this, or should I create a separate issue for it?

[Gozala]

Is analog to HTTP's Range requests tracked under this, or should I create a separate issue for it?

Let's track it here for now, depending how it will pan out I might need separate one, but for now it's good. One thing that could help is an example that produces Range requests that are disregarded.

[lidel]

@Gozala I believe seeking with <video> is the most popular use for range requests right now.

I created a small sandbox illustrating current problems, more details in README at:

• lidel/libdweb/tree/video-range-use-case-demo

[Gozala]

Migrate to https://bugzilla.mozilla.org/show_bug.cgi?id=1572215