Remko Bolt <[EMAIL PROTECTED]> writes: > Would you consider supporting negative keys to facilitate sorting by the last > column with --key=-1?
That makes sense to me, yes. I suppose key 0 could be the entire line, so key -0 would be equivalent to key 0. One other idle thought: currently 'sort -k 9999999999999999999999999999' complains, but there's no real reason for it to complain; it could just silently ignore the key. Even if infinite-precision arithmetic were supported that key can't possibly contribute to the comparison, since the input record can't possibly contain that many keys due to size_t limitations. _______________________________________________ Bug-coreutils mailing list [email protected] http://lists.gnu.org/mailman/listinfo/bug-coreutils
