Make Xpath accept slightly malformed xml (like html)

[AlfredBerg]

Currently if the xml that a XPath expression is ran against is not perfectly formatted without any errors or warnings you will get Invalid input XML.. The xmldom package is still able to run XPath expressions even if the XML is not perfectly formatted (like most html documents).

Example:
A simple html document that does not close all tags https://gchq.github.io/CyberChef/#recipe=XPath_expression('//meta','%5C%5Cn')&input=CjwhZG9jdHlwZSBodG1sPgo8aHRtbCBsYW5nPSJlbiI%2BCiAgPGhlYWQ%2BCiAgICA8bWV0YSBjaGFyc2V0PSJ1dGYtOCI%2BCiAgPC9oZWFkPgogIDxib2R5PgogIDwvYm9keT4KPC9odG1sPgo

This pull request makes it so the XPath expression will work in the example.

[CLAassistant]

All committers have signed the CLA.

[a3957273]

Oh boy is this an old change, but looks good to me! Thanks for the change. :)