> After receiving at least 200 404 reports from my website 
I 
> can safely say that the method HTTrack uses to determine 
> links from JavaScript files is flawed.

The method is not perfect, and often fails with "complex" 
javascript. There is no way to perfectly detect "javascript 
links", even using javascript engines. This is a limit in 
all offline browsers, and I don't think it will be solved 
one day.

(Note that 404 errors should not consumme too much 
bandwidth, as only the return status and a small page is 
returned, generally)