Last modified: 2013-12-05 19:31:06 UTC
Parsoid: echo "{{گفتاورد بزرگ}}" | tests/parse.js --prefix=fawiki emits: <body data-parsoid='{"dsr":[0,17,0,0]}'><p about="#mwt1" typeof="mw:Transclusion" data-mw='{"parts":[{"template":{"target":{"wt":"گفتاورد بزرگ","href":"./الگو:گفتاورد_بزرگ"},"params":{},"i":0}}]}' data-parsoid='{"dsr":[0,16,null,null],"pi":[[]]}'><blockquote|> </p> </body> The <blockquote|> business is totally bogus. It's not present in the output of the PHP parser as far as I can tell. This is an example from https://fa.wikipedia.org/wiki/%D8%A2%D8%B1%D8%A7%D9%85%DA%AF%D8%A7%D9%87_%DA%A9%D9%88%D8%B1%D9%88%D8%B4_%D8%A8%D8%B2%D8%B1%DA%AF#cite_ref-22 and you can see the bogus 'blockquote' stuff in http://parsoid-lb.eqiad.wikimedia.org/fawiki/%D8%A2%D8%B1%D8%A7%D9%85%DA%AF%D8%A7%D9%87_%DA%A9%D9%88%D8%B1%D9%88%D8%B4_%D8%A8%D8%B2%D8%B1%DA%AF?oldid=11247123#cite_ref-22-0
$ echo '<blockquote|>a</blockquote>' | php maintenance/parse.php <blockquote>a</blockquote> $ echo '<blockquote|>a</blockquote>' | tests/parse.js <body data-parsoid='{"dsr":[0,28,0,0]}'><blockquote|>a </body>
We could try to make our generic_newline_attribute production slightly more tolerant of broken wikitext like this. If we can achieve this with limited effort and without breaking the parsing of other content then it would be a good thing to do. It would also not hurt to fix up the wikitext, possibly with help from our end (bug 46705).
