Hi Chris, On Wed, 2007-01-24 at 14:22 +0800, Chris Wang wrote: > Laca: > May be there is some misunderstanding, the audio CD cannot be played > without this patch on Sparc. the use of environment > variable give the minor user who use big endian CDROM(very rare) to > switch this patch.
You're right, that wasn't clear. The new patch (without the env variable) looks reasonable and upstreamable. Please file a bugzilla bug. Thanks, Laca > For X86 user, they are using little endian(most case) CD_ROM and little > endian processor, so this problem is not exist. > the problem only exist when the endianness of CDROM and processor is > different, for most of the case little endian CDROM and big Endian > processor (Sparc ). > > In order for major part of user can use the application properly, I do > believe this patch is necessary. > > Chris > > Laszlo (Laca) Peter ??????: > > Well, if this is a "temporary fix" that gives certain users who > > use a "very rare" cdrom drive "some ability" to play audio CDs, > > then I don't think we should do this in HEAD and certainly not > > in the stable branch. > > > > Laca > > > > On Wed, 2007-01-24 at 12:14 +0800, Brian Cameron wrote: > > > >> Laca: > >> > >> > >>> On Wed, 2007-01-24 at 11:41 +0800, C Wang wrote: > >>> > >>>> This patch solved this problem while leave the user to switch off the > >>>> patch if the CDROM is big endian( very rare ) by setting the > >>>> environment > >>>> variable CDROM_BYTE_ORDER. > >>>> > >>> Is there a way to detect this automatically? > >>> Setting an env variable is not a nice way to control this, I would think > >>> that most people probably start sound-juicer from the Launch menu and > >>> not from the command line. It's also a new interface that should > >>> be ARC-reviewed and documented. > >>> > >> I agree. This should be viewed as a temporary fix. Unfortunately > >> libcdio doesn't give any interface for determining endian-ness of the > >> drive. The cdparanoia (a Linux-only library similar to libcdio) does > >> have some code, but it isn't well written and uses statistics to > >> "guess" the endianness of the drive. Not sure this code is good to > >> use here. > >> > >> Until we have some code to do endianness detection, this at least > >> gives users some ability to configure GStreamer so that it will > >> work on drives with the problem. > >> > >> Brian > >> > >> > > > > > >
