rm -f $outfile.inc
-if test -n "$excfile"; then
+if test -n "$excfile" -a -f "$excfile"; then
process_includes $excfile $outfile.exc
fi
fi
if test -f $outfile.exc; then
+ # So what we're doing below with uniq -u is that we take
+ # lines that have not been duplicated. This computes the
+ # symmetric difference between the files. This is not
+ # what we want. If a file is in the excludes but not in
+ # the sources, we want that file not to show up. By duplicating the
+ # excludes, we ensure that we won't end up in this failure state.
sort -u $outfile.exc > $outfile.exc_s
- rm -f $outfile.exc
- sort -m $outfile.inc_s $outfile.exc_s | uniq -u > $outfile
- rm -f $outfile.inc_s $outfile.exc_s
+
+ # Duplicate excludes
+ cat $outfile.exc_s >> $outfile.exc_s_dup
+ cat $outfile.exc_s >> $outfile.exc_s_dup
+
+ rm -f $outfile.exc $outfile.exc_s
+
+ cat $outfile.inc_s $outfile.exc_s_dup | sort | uniq -u > $outfile
+ rm -f $outfile.inc_s $outfile.exc_s_dup
else
mv $outfile.inc_s $outfile
fi