But we don't need to convert it to UTF8, who said that? ;) We ned to have some encoding which is recognized by xsltproc first. If windows-1255 is the same as iso-8859-8-i, then this is simply a matter of find/replace in the XML files to replace the encoding names, and then xsltproc will hopefully work.
This is what I really don't understand at all... Why does it need to be converted, at all? The hebrew as it is in the original translated files is already in the form that if will be displayed in the browser, will look fine! I mean - don't change anything and it works. As long as you output the relevant HTML tags in the final HTML output (that would be: <html dir="rtl"> and <meta charset="iso-8859-8-i">) - it will look OK.
If xsltproc is just trying to be naughty, I think that even if you put
there (at the encoding) the value of iso-8859-1 (english) - the hebrew
will be displayed fine at the browser. I can't really check because I
really don't know how, but if you want to make such a file and send it
to me, I can see if it looks fine :)
Please read again what I have written. I am not talking about iso-8859-1. Where does it come to your mind??
Goba