Closed jfolz closed 11 years ago
That, uh, shouldn't be happening at all. I'll look.
As expected, parsing happens correctly using the Python tokenizer. Looking into what's up with the C tokenizer.
Edit: Think I know what's going on. Edit 2: Solution ready; seems to work. Testing further.
Should be resolved now. As usual, the develop branch contains a work-in-progress C tokenizer that may be buggy and prone to these issues. Please report them whenever they come up!
I've come across this little gem here:
It's from here: http://de.wikipedia.org/w/index.php?title=Strausberger_Platz&oldid=112496475
As soon as the line
Bild=Strausberger Platz Berlin April 2006 109.jpg|miniatur
is added to the template, it is rejected and a Text node is created instead. Even though this markup doesn't make sense, I would expect a Template node with a value-less parameter named "miniatur" instead of a Text node.