[XML-SIG] Big Bug? (was:Pretty-printing DOM trees)

[Lars Marius Garshol]

* Christian Tismer
| 
| Maybe xmlproc should be a little more forgiving for this case and
| not skip beyond ">" but just skip (or repair) the attribute.

* Greg Stein
| 
| It should *NOT* repair the attribute. That will simply encourage
| poor XML authoring. It should report the error properly (or,
| alternatively, the error should be responded to properly).

The error is reported properly as it is and the attribute is not
repaired, but subsequent data events are wrong. That's now fixed (data
events, not the attribute), but the question remains whether the
parser should follow the XML recommendation and stop reporting data
events after a well-formedness bug.

I'm inclined to make that default behaviour, but behaviour it is
possible to turn off. Opinions are welcome.

--Lars M.