Re-enable custom target parser support

[abelcheung]

This is most likely achievable by defining a stub-only CustomTargetParser, derived from XMLParser.__new__() whenever target argument is specified. More work is involved besides the stub-only class:

• C14NWriterTarget and TreeBuilder need to inherit from ParserTarget and match its interface
• Create wiki page describing how to use custome parser target in fully annotated source code
  • need to enlist all methods available in target object
• Tie up loose ends of annotation-only ParserTarget class
  • arguments in target object methods need to be positional only, in order to avoid pyright warning about inconsistent param name
• All methods and functions across the whole types-lxml need to add @overload to handle such custom parser
  • HTML()
  • XML()
  • _ElementTree.parse()
  • ElementTree()
  • fromstring()
  • fromstringlist()
  • parse()

It only works for pyright; mypy ignores any @overload which __new__() returns anything not from originated class or its subclass. Until python/mypy#15182 is fixed, instruct test suite to ignore mypy completely.

---

Boilerplate code for parser target might look like:

if TYPE_CHECKING:
    from lxml.etree import ParserTarget
else:
    class ParserTarget:
        __class_getitem__ = classmethod(GenericAlias)

class MyTarget(ParserTarget[int]):
    def close(self) -> int:
        ...  # do your stuff

parser = XMLParser(target=MyTarget())
reveal_type(parser)  # CustomeParserTarget[int] for pyright, XMLParser for runtime
with ... as file:
    result = XML(file.read(), parser)
reveal_type(result)  # int for both pyright and runtime