http://bugzilla.novell.com/show_bug.cgi?id=480178
http://bugzilla.novell.com/show_bug.cgi?id=480178#c17 --- Comment #17 from Damien Diederen <[email protected]> 2010-04-21 19:40:47 UTC --- Hi Paolo, (In reply to comment #16) > I was playing with a bi-level table compression myself a few months > ago, so the general approach is fine by me. Okay. > On the specific implementation, I'm not sure some of the additional complexity > in your changes is worth it. Let's consider a 256 byte page size. Any category > lookup could be done with: > char_data [char_start [val >> 8] + (val & Oxff)] > this is more compact and in most cases should be better than the branchy code > in your patch. Care to try that out or did you already test something similar? Possibly. This implementation is a (more or less) direct port of GLib's, which itself comes from libunicode, and I must admit I haven't had time to research the history of that solution, nor to explore alternative ones. This is definitely worth trying and measuring, though. I will look into it before submitting an updated series. > As for multiple versions of the data, we likely just want to use the > latest, but once we have numbers about the cost of this we could > reconsider. Okay; I will focus on getting numbers for the simple lookup technique first. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the QA contact for the bug. You are the assignee for the bug. _______________________________________________ mono-bugs maillist - [email protected] http://lists.ximian.com/mailman/listinfo/mono-bugs
