Fix ElementTree serialization to HTML

[serhiy-storchaka]

Bug report

There are several bugs which prevent generation of correct HTML:

• Comment and "processing instruction" content should not be escaped. Similar bug was fixed for XML in ElementTree ProcessingInstruction uses character entities in content #46995.
• There are more elements besides <script> and <style> whose content should not be escaped.
• <plaintext> does not have the closing tag, and its content should not be escaped.
• Empty attributes (with value None) are not supported.

Linked PRs

• gh-149489: Fix ElementTree serialization to HTML #149490

[serhiy-storchaka]

There is a question what to do with <noscript>. It's syntax depends on whether scripting was enabled.

[serhiy-storchaka]

Well, it is only issue for the parser. The generator should escape it's content, and allow subelements.