yorkz1994

No, `PartialEq` is not the same as `PartialOrd`. Here: https://docs.rs/tracing-core/0.1.32/tracing_core/metadata/struct.Level.html It clearly said: ``` use tracing_core::Level; assert!(Level::TRACE > Level::DEBUG); assert!(Level::ERROR < Level::WARN); assert!(Level::INFO Option { Some(filter_as_usize(&other.0).cmp(&(self.0 as usize))) } ......

Since this is a document bug, could you please correct it. That is the reason why I opened this issue.

I think the failed test is because we did the line end normalization after the unescape, but unescape can contain `\r`, so these new `\r`s should not be normalized. We...

> Good. But I would also like to see the tests when use API of `Attribute` and `BytesText`, which are listed [here](https://docs.rs/quick-xml/latest/quick_xml/?search=unescape). Sorry, don't know what do you want me...

Like I said, I am not familiar with XML specification, so I don't know where is the proper places to put `\r` in it and what should be expected after...

Could you provide such input data. I don't want to create such invalid XML. Normally I can think is that the `\r` will only appear in `BytesText` due to line...

[f970370](https://github.com/tafia/quick-xml/pull/807/commits/f970370ff64ee431094ad667620df8211a5e7656) This commit is due to forgot to normalize if input has nothing to unescape. I add a case from your input, don't know the result is your expected: ```rust...

The failed test is because there is `\r` in the roundtrip test. It is impossible to have equal serialize and deserialize due to line end normalization. Therefore I have to...

I did more line ends normalization. Also changed the normalization implementation to use iterator to avoid extra allocation during normalization.

@jmcnamara Thanks. This information is very useful. I check the code, it seems only `_x00HH_` literals are escaped. If other valid `_xHHHH_` literals are skipped, then when doing read, excel...