On Mon, 27 Oct 2008, David Roundy wrote: >> Sat Oct 25 13:18:57 BST 2008 Ganesh Sittampalam <[EMAIL PROTECTED]> >> * accumulate hash strictly in hash/hashPS > > I've applied the above. > >> Sat Oct 25 13:23:31 BST 2008 Ganesh Sittampalam <[EMAIL PROTECTED]> >> * just hash the last 20 characters in LCS > > Is there a particular bug that this fixes? It sounds like the sort of > change that could easily be a pessimization (due to increased hash > collisions), so I'd like to see some sort of numbers to justify it.
It was intended as a "belt and braces" fix for 1017, as well as something that's likely to be a good idea in its own right. I'll see if I can provide some numbers while benchmarking some other work I did on slurpies, but I doubt it's that important or worth worrying about. I think increased hash collisions in the general case are highly unlikely, the risk would be for some weird kinds of data. >> Sat Oct 25 13:27:54 BST 2008 Ganesh Sittampalam <[EMAIL PROTECTED]> >> * issue 1017 now fixed > > When was this fixed? It still fails for me? Either accumulating the hash strictly, or limiting the characters hashed, should fix it. Ganesh _______________________________________________ darcs-users mailing list [email protected] http://lists.osuosl.org/mailman/listinfo/darcs-users
