[DiapDealer]

I'm guessing it's some of my shoddy regex work at lines 70-73 of mobi_html.py.
I'm not entirely certain of Python's default lazy/greedy settings for it's regex engine, but it seems to me that something similar to (line 71):

link_pattern = re.compile(r'''<a(.*?)filepos=['"]{0,1}0*(\d+)['"]{0,1}(.*?)>''', re.IGNORECASE)

and consequently (line 73):

srctext = link_pattern.sub(r'''<a\1href="#filepos\2"\3>''', srctext)

might do the trick (unless that id attribute itself needs additional massaging, that is).

Spoiler:

Or blow it up... my regex-fu is not mighty.