The implementation of DependSet was O(n^2) because of a double nested loop. In other words, 1 million comparisons were being made for 1000 files (if everything is up to date, which is a common case). I reduced it to O(n), so just about 1000 comparisons for 1000 files. It's MUCH faster now.
Thanks Mohammad (See attached file: DependSet.diff)
DependSet.diff
Description: Binary data
-- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
