Hi, > > http://tug.org/svn/texlive/trunk/Master/support/tests/ Hmm.. this approach is in fact pretty good in that it will ignore details of variant parts within generated PDF files (like the font information, and time), and is probably a bit more robust than using 'diff'. As Frank pointed out, it does lack checking for links and stuff, so that will need to be checked in other ways. regards, junichi -- dancer@{debian.org,netfort.gr.jp} Debian Project