@j4k0xb

Thank you for your contribution, great work!

That said, I have two concerns.

First, this change introduces a dependency on the DOM for mutf-8. I aim to keep the dependencies of mutf-8 as minimal as possible. While I haven't fully thought through a solution yet, one idea could be to split the module into mutf-8 and mutf-8-streams.

The second concern is a bit more significant. There might be an issue with decoding multibyte characters in this implementation. The problem occurs when a chunk is split in the middle of a byte sequence representing a multibyte character. TextDecoder has a stream mode that handles incomplete byte sequences at the end of one chunk by combining them with the start of the next for decoding. Unfortunately, MUtf8Decoder does not yet implement stream mode.

For these reasons, I’ll need to close this pull request.

However, I encourage you to keep contributing, and I’m open to discussing possible solutions to these issues!