Sizes differing too much

[theamato]

Hi,
I'm trying to align an Arabic-English text corpus by using the following command: src/hunalign/hunalign -text -hand=hand_align_file dict.txt ar.txt en.txt. I don't have a dictionary, so dict.txt is an empty text file. When I run it I get this error:

Reading dictionary...
20 hungarian sentences read.
0 english sentences read.
Sizes differing too much. Ignoring files to avoid a rare loop bug.

Am I missing something? Why aren't the english sentences read? Should I provide language codes when working with a different language than hungarian?