Eric Siegerman writes: > > True. Sorting directory contents in "update", for example. But > I'd have thought that, except for *huge* directories (single > directories, not subtrees), the (in-memory) sort would be dwarfed > by all that O(n) file I/O.
Anything that uses recursion sorts directory contents. I was thinking of that kind of overhead, which is, most likely, O(n log n), but you're right that other operations are dominated by file transfers. Those operations are O(s) where s is the total size transfered which is not really related to the number of files in the repository, the working directory, or even being processed by the command. I think the bottom line is that CVS performance is hard to characterize, but the number of files in the repository as a whole generally doesn't matter very much. -Larry Jones Like I'm going to get any sleep NOW. -- Calvin _______________________________________________ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
