ZegnatThat matches my recollection. I would definitely consider a parser to pass a test if their output is completely correct except for the />. Though I would then appreciate 1 test that is not normalised so it could still be flagged whether the parser follows spec or not.
