Parser unescapes &lt;, nodeToString leaves the < as is

[malidukica]

Hello @hecrj, and thank you for the amazing library and work!

The issue that we came across is related to a roundtrip that we do, html string > html parser nodes > html string, since we need to do some adjustments to the html string.

We found out that the parser unescapes a < to <, and stringifying it again leaves < as is (which in turn causes a parsing error when we try to parse the html string anew).

The example is given in the following ellie:

https://ellie-app.com/ndLhBgPfqvDa1

Are there any special considerations for this behaviour? (I'm not a huge expert in html sanitization and escaping)