Matthias Ringwald a écrit : > Hi Linus, hi Mark. > > Besides stressing Alexander's suggestion on better using the tracker for > discussing .info submissions than sending inline text (which often > causes trouble), I'd like to learn about the required patches. If they > are not fink specific but just required on the mac, we should try to > push them upstream to xine-devel (which is appreciates most of the times).
OK, it's in the tracker now, with updated info-file. Sorry for the bug. > Did you read my 2 previous postings to the list concerning xine lib on > x86-32 with XCode 3.0? The MacBook Air has a Core 2 Duo which AFAIU is a > 64-bit CPU (please correct me if I'm wrong). I can try to see if your > .info/.patch compiles on a MacBook x86-32 with XCode 3.0 next week. Hmm, I'm still not clear whether this is 32-bit or 64-bit. Core2Duo is 64-capable, but sizeof( int* ) is still 4, so this seems to be 32-bit. > In the info file, xine lib 1.1.10.1 is used which is already superseded > by 1.1.11, which I belief will be replaced by 1.1.11.1 pretty soon, as > another security bug was discovered recently. But this shouldn't make > any difference to the .info file besides the source code url. > In the info file %v can be used, e.g., in the Source tag to specify > xine-lib-%v.tar.bz instead of repeating the version number. My bug, corrected version is in the tracker. > I'm not sure if most of the patches from the PatchScript are still (it > looks like the come from the current libxine 1.1.2 package) needed. As I > said, I can build xine lib 1.1.11 on G4, G5 without any patches, and on > X86-32 at least with XCode 3.1 when disabling tomsmocomp. > Maybe we can start without patches and add them as needed. I don't have XCode 3.1. When I try to compile with the included ffmpeg, I still get the dreaded ld: absolute addressing (perhaps -mdynamic-no-pic) used in _postProcess_MMX from .libs/xineplug_post_planar.lax/libpostprocess.a/postprocess.o not allowed in slidable image collect2: ld returned 1 exit status so compiling with --with-external-ffmpeg it goes through (when disabling dscaler_greedyh and tomsmocomp). I put a new .info and .patch to the tracker, a bit shorter now. > The patch disables the dscaler_greedyh. I think the problem with its asm > implementation was already fixed in xine-lib and should work now on mac > x86. disabling the dscaler_tomsmocomp is probably required. Would it be > better to keep the method but make it a No-op, such that the > deinterlacer can be selected but it's doing nothing? Might that be a problem with GUIs for Xine? Else I'd say to just delete it. > Linux, for xine-lib not working on your machine. can you try to > download, and install xine-lib without fink into the default location > (/usr/local/..) and then try the muxine example again? Could you erase > xine lib before from your machine too? uninstall the ffmpeg packages > from fink, clean up your /usr/local/lib and /usr/local/include. maybe > some files got mixed up. Yep, now it works fine, thanks! But only if I install it in the root, with fink it still doesn't work :( Linus ------------------------------------------------------------------------- Check out the new SourceForge.net Marketplace. It's the best place to buy or sell services for just about anything Open Source. http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace _______________________________________________ Fink-devel mailing list Fink-devel@lists.sourceforge.net http://news.gmane.org/gmane.os.apple.fink.devel
