MXParser improve error reporting when parsing large char entities and invalid encoding declarations

[belingueres]

Discovered with the help of https://github.com/belingueres/xml-conformance

MXParser fails on parts of the W3C XML Conformance Test Suite.

In this issue, it is failing 001.xml to 004.xml, and 007.xml to 009.xml from "Bjoern Hoehrmann via HST 2013-09-18" (see https://github.com/belingueres/xml-conformance/blob/main/src/test/resources/xmlconf/eduni/misc/ht-bh.xml) regarding the parsing of large character entities and encoding in the declaration section of the xml file.
Tests 005.xml and 006.xml were ignored as MXParser do not support DOCDECL parsing.

[michael-o]

Fixed with 1e18ddc.