Move SourceBufferWriteHead to Cobalt module

[osagie98]

1. Moves the SourceBuffer.writeHead implementation from the mediasource module to the Cobalt module.
2. Renames source_buffer_write_head.idl to source_buffer_extensions.idl.

b/395658866

[osagie98]

This is a quick look at how it may look when we move custom interface extensions to the Cobalt module. This can't build due to a dependency issue. However, I don't think we would need to deal with Mojo here. This also concerns #4837, which will need to do something similar once the process is defined

[hlwarriner]

This looks like a good change.

Re Mojo, does SourceBufferWriteHead.writeHead() end up calling a Starboard API at some point in the call stack?

[osagie98]

It doesn't, writeHead goes as far as the ChunkDemuxerStream in the chromium media stack

[osagie98]

Pending verification with the web platform test. This also needs a unit test - that can be addressed in a separate PR.

[sideb0ard]

this looks good, I copied some of this structure for #4837 and it works well.

[osagie98]

As WPTs aren't currently working, I'll verify them later. Also, ChunkDemuxerTests are broken/unverified and need extra work to get going. I think it should be fine to merge as is and verify through tests later. I'll still verify using a Cobalt demo before merging.

[hlwarriner]

As WPTs aren't currently working, I'll verify them later. Also, ChunkDemuxerTests are broken/unverified and need extra work to get going. I think it should be fine to merge as is and verify through tests later. I'll still verify using a Cobalt demo before merging.

Sounds good to me if you've tested locally and this works as expected. It would probably be good to have a bug to track the WPT; you could mark it as blocked by b/392940729.

[osagie98]

Verified writeHead works with a demo

[yell0wd0g]

Moving files LGTM!