>On what level checks httrack if the file already has been 
>downloaded - on filesystem base or on link base? I 
>compared 

Case insensitive (except for the hostname), full link 
(including query string) based comparison. Duplicate 
entries are detected, and collisions too (index.html and 
Index.html, for example)

> I just tried to exclude *pdf*, but httrack ignores this. 
> what's wrong with my call?
?? Darn?! -*pdf* (not -*pdf) should work, and should refuse 
all links containing "pdf" inside it.
You can try -*pdfdownload.php*, BUT -*pdf* should 
definitely work (you can also try -*?*pdf*)

Are you sure the option isn't interpreted by the shell or 
something (check in hts-cache/doit.log)?