First of all, thank you a lot for the fb22lrf converter! I like it a lot.
And since revision 1655 of src/calibre/ebooks/metadata/fb2.py, it is almost perfect.
But some small bugs are still present and it would be nice to get them also fixed.
F.e. I am wondering why notes are always points to the previous one. It looks like a general problem because generated with --verbose /tmp/html2lrf-verbose.html file is correct.
Also the converter shows an error 'Could not find image' every time when fb2 has images.
The book
http://lib.rus.ec/b/72606/download may be used as a reference for the both problems.