Potential Bug for UTF-8 BOM strip

[zhongyi51]

According to the section 9.2.6 of SSE standard:

The UTF-8 decode algorithm strips one leading UTF-8 Byte Order Mark (BOM), if any.

However eventsource-client does not handle the leading BOM head now.

To fix this bug, we can introduce a flag first_chunk in EventParser. The parser will check and skip the BOM during parsing the first chunk.

If it is OK, I'm happy to submit a PR for this~

[keelerm84]

Thank you for raising this issue. Yes, please feel free to open a PR.

we can introduce a flag first_chunk

By flag here, do you mean an option to enable or disable this BOM stripping behavior? If so, I don't think you need to bother with that. The spec states it should be stripping this, so it just seems like a bug we should address.

Thanks again!