Hello ! I'm on an iBook G3 machine (ref M8600LL/A on Apple website), which should be a non Altivec one if I understood well (?).
I tried your test and the result is "Instruction illégale". Versions : - gstreamer-base and -tools : 0.10.10-2 - kernel : linux-image-2.6.18 (locally compiled from Debian source with Debian conf) I then downloaded the new libvisual you pointed. The result has change : "ERREUR: le tube n'a pas pu etre construit: tube vide non autorisé." (The pipe couldn't be constructed, empty pipe forbidden). What do you want me to do next ? Hoping it could help you... Bertrand ------------------------------------ Original Message Follows : De: Sjoerd Simons <[EMAIL PROTECTED]> Répondre à: [EMAIL PROTECTED] À: [EMAIL PROTECTED], debian-powerpc@lists.debian.org Cc: [EMAIL PROTECTED], [EMAIL PROTECTED], [EMAIL PROTECTED], [EMAIL PROTECTED] Sujet: G3 (or other non-altivec machines) testers sought Date: Sun, 14 Jan 2007 23:08:03 +0100 Client de messagerie: Mutt/1.5.13 (2006-08-11) Hi, Gstreamer currently has a release critical bug that seems to be only reproducable on powerpc machines without altivec support and even then not always. The best hint to what the problem might be seems to indicate there is an issue in the libvisual altivec detection code[0].. Also for some reporters the problem went away (!?). Google revealed that pld-linux re-enabled the altivec detection because it works ``on fixed kernel''. So it might be related to specific kernel versions. So what i would like people to do (Only on machine without altivec!): 0: Install gstreamer 0.10 (at least gstreamer0.10-plugins-base and gstreamer0.10-tools from etch) 1: Run: rm ~/.gstreamer-0.10/registry.*.xml gst-launch-0.10 --gst-disable-registry-fork 2a: If it doesn't crash: - please try on an older kernel if possible 2b: If it does crash: - please try on a newer (debian) kernel if possible - if that doesn't help try the libvisual packages at: http://people.debian.org/~sjoerd/libvisual/ 3: Thanks for testing! Please mail your experiences, kernel versions, machines types etc. Sjoerd 0: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=404876;msg=88 -- Don't kid yourself. Little is relevant, and nothing lasts forever.