On Thu, Feb 27, 2003 at 06:09:18AM -0800, [EMAIL PROTECTED] wrote: > On Thu, Feb 27, 2003 at 09:04:59AM -0500, Barrie Slaymaker wrote: > > On Thu, Feb 27, 2003 at 01:44:50AM -0800, [EMAIL PROTECTED] wrote: > > The above is true, and beneficial, but your post mentioned something > about compare() being run less often and less screen clutter. This would > seem to mean compare is run *more* often since the loop is skipped > less often.
The primary effect of the patch is to pay closer attention to the source and destination mtimes and not run compare() if the mtimes are equal; the file size comparison is there just in case somebody's been playing with tar or cpio or even running scripts that move so fast that they achieve subsecond resolution on changes (something I've been tripping over a lot with CVS lately) and happens to get identical mtimes. I didn't think about skipping the compare() if the sizes differ, but that makes sense and would probably skip most of the compare() calls from pm_to_blib(), which might be noticable on larger lib/... trees or slower machines. Good catch. - Barrie
