Amara3.xml

* HTML5 handling fixed
* Python 3.9+ support

Fixed:

* Escaping problems in `amara3.uxml.writer`
* Bugs in `boolean()` & `number()` XPath functions
* Other minor bugs

Fixed:

* Obsolete use of `treesequence` in `microx`

Added:

* Implement uXPath operators or/and

Changed:

* Better modularize & name iterated tree parser modules

Fixed:

* Iterated tree parser modules
* uXPath node tests
* element.xml_write()
* Test cases

Notes:

* Warning: html5lib is the underlying parser, so unfortunately Python 3.8 support will be broken until we have an [upstream fix](https://github.com/html5lib/html5lib-python/issues/419).

Added:

* Namespace packaging
* Add `--html` option and `--find-text` command (with `--show-attrs` option) to `microx` cmdline.

Changed:

* Nothing of note

Fixed:

* Packaging

* Fix retrieval of virtual root nodes, and thus absolute paths
* Fix amara3.uxml.isxml()
* Fix node() and text() tests in XPath
* Implement amara3.uxml.uxpath.qqery()
* Other fixes.