#21 Less-Strict XML Parsing

[Morbus Iff]

Previous to the 0.93.x branch of AmphetaDesk, we'd attempt
to fix broken XML by doing entity encodings, starting
whitespace removal, etc.. In 0.93+, we moved to a totally
strict XML parsing edict. AmphetaDesk should revert back to
a default of soft parsing, with the capability to turn on strict.
Likewise, anything that fails a strict parse, but not a soft,
should be alerted in the primary page. The soft parsing code
for this is hanging around in the CVS revision history, just a
matter of pulling it out.