Bináris wikiposta@gmail.com:

Try this first in replacements:

ur'(?m)^== *([a-zA-Z ]+) *==\n+ etc. m for multiline (sometimes ?s works instead, I just use it without understanding why), and space* in case there are spaces between == marks.

Sorawee Porncharoenwase nullzero.free@gmail.com:

@Michael Hamm: Beside Bináris's suggestion, re.escape should be used only on pattern parameter, not replacing text; use match.group(1) instead of re.escape(match.group(1))

Thank you both very much. Probably (I won't be able to test it until later today) these were the problem -- or part of it.

Michael Hamm