I've written a small script that traverses a directory tree and does some inplace editing on HTML files. Curiously, when I use find to do a breadth-first traversal the script seems to get stuck when it encounters a directory. With finddepth it's as fast as one would expect. BTW, I tried to have perl do the inplace editing, but I couldn't get the $^I stuff to work and I don't know how to set the -i command line switch for a droplet. Here's the interesting part of the script, maybe there's something I've done wrong? sub wanted { if (-f $File::Find::name) { process($File::Find::name); } } sub process { my $path = shift(@_); return unless $path =~ /\.(html|htm)$/i; print STDERR "Processing $path\n"; $tmpfile = $path . '~'; rename($path, $tmpfile); open(INFILE, $tmpfile) || die "Couldn't open file $tmpfile: $!\n"; open(TMPFILE, ">$path") || die "Couldn't open file $path: $!\n"; while (<INFILE>) { s/<!--[^>]*-->//ig; s/<META[^>]*>//ig; s/<x-sas-window[^>]*>//ig; print TMPFILE; } close(INFILE); close(TMPFILE); unlink($tmpfile); } Michael --- Michael Schuerig Papa, draussen steht ein Philosoph. mailto:uzs90z@uni-bonn.de Sag ihm, wir kaufen nichts. http://www.uni-bonn.de/~uzs90z/