> <img src="image.php?id=23423" />
> where image.php is some script that will return an
> image depending on the id.
> Now, it seems that HTTrack will see this link and
> download it.  The problem is that php automatically
> sends the http header as "text/html" unless you
> change it explicitly, which the author of the site
> in question did not do.

Then httrack won't see it as image.. this is the same with Mozilla, for
example (the image won't be properly displayed with Mozilla or Firefox AFAIK)

> Warning: File not parsed, looks like binary ...
> Is there a switch somewhere that will make HTTrack
> try to figure out (best guess) the MIME type NOT
> based on its http header?
No, because even if recent httrack releases (3.40) now wait for remote headers
before naming the destination file, the content itself is downloaded in
bakground.