Daniel Alley

> I'm not sure, that we should remove StartText event, because it still can be useful even if will not contain BOM. The text before the declaration is not considered...

>Because almost all supported encodings are one-byte encodings, we could create XML files with all possible symbols (probably except that explicitly forbidden by the XML standard). The national symbols should...

> detecting encoding (I found that we do not do that correctly, I'll plan to look at this before release, I already made some investigations and improvements, but not yet...

The basic functionality is pretty much working. TODO: - [ ] encoding detection regression for UTF-16 (failing tests) - [ ] async support in `Utf8BytesReader` - [ ] UTF-8 validation...

Status update: I'm still working on this, but making PRs against https://github.com/BurntSushi/encoding_rs_io/

And also https://github.com/hsivonen/encoding_rs/issues/88 I'm not sure how long I'm going to be blocked on that, so if there is a portion of the work you think could be split off...

So, as a practical matter I now think we need to strip the BOM in all situations, no matter what. Not because there's anything invalid about it at a purely...

@mingun I'm just going to rip out any changes to the serde code and do that absolutely last, so I don't need to keep redoing it It's been a month...

Is this related to https://github.com/tafia/quick-xml/pull/337?

I've been thinking about this a bit lately (and writing my thoughts on this [PR](https://github.com/tafia/quick-xml/pull/415#issuecomment-1180691911=)) so I will copy them here where they are more relevant. I agree with the...