Last modified: 2014-08-20 22:34:45 UTC
Some articles contain de-facto HTML-style comments like <!--- foo ----> (with more than two dashes). Parsoid passes these unchanged into the output XML, which is then invalid. Failure example: $ wget -q 'http://parsoid.wmflabs.org/enwiki/Bratislava?oldid=617085374' -O - | xmlwf STDIN:6:3: not well-formed (invalid token)" Non-failure example: $ wget -q 'http://parsoid.wmflabs.org/enwiki/Bratislava?oldid=617286772' -O - | xmlwf