Hi, Am Mittwoch, den 24.03.2010, 16:03 +0000 schrieb brian m. carlson: > On Tue, Mar 16, 2010 at 10:54:32AM +0100, Joachim Breitner wrote: > > Besides that, I’d be glad if someone else looks over the code, maybe I’m > > missing something. It can be found at > > http://git.nomeata.de/?p=unicode-screensaver.git;a=blob;f=unicode.c;hb=HEAD > > My screen is set to blank after a short amount of time after the > screensaver comes on, so I usually only see a few characters per run. > > I took the version in the archive (0.2) and looked at utils/yarandom.c. > Apparently, the algorithm implemented is not the algorithm it claims. I > looked at my copy of Knuth and there are some differences. Whether > these affect the generator's effectiveness for this matter, I don't > know. > > I patched the screensaver to dump some data for each character it > displays and then ran it thrice for 11 characters per run. Those are > attached as dump1, dump2, and dump3. The patch that I used to generate > this is also attached. Note that where the same characters appear, the > same numbers are generated (look, for example, at the sequence starting > with U+A710 and continuing through U+1503). > > My guess, just from looking at the code, is that perhaps the seeding > process is not effective. It seems like it modifies the upper bits a > great deal, but has little to no effect on the lower bits. I'll take a > look in more detail shortly.
Given that I take the code directly from the xscreensaver distribution, I’d be surprised if the bug would be in yarandom.c. If it is so, it should probably be fixed there. I’m excited what you’ll find out! Greetings, Joachim -- Joachim "nomeata" Breitner Debian Developer [email protected] | ICQ# 74513189 | GPG-Keyid: 4743206C JID: [email protected] | http://people.debian.org/~nomeata
signature.asc
Description: This is a digitally signed message part
