On 10/04/2010 09:39 PM, Harry van der Wolf wrote:

2010/10/4 Nikos Chantziaras <[email protected] <mailto:[email protected]>>

    On 10/04/2010 09:07 PM, Brandon S Allbery KF8NH wrote:

        On 10/3/10 21:21 , Nikos Chantziaras wrote:

            Thanks, Ryan.  I suppose I will need to find someone who has
            a 10.5
            installation to do a sanity check for me before I release
            any binaries
            tagged with "OS X 10.5 and newer".


        Distributing binaries built via MacPorts tends to be a bad idea:
          not only
        do you need to make sure to distribute the files from any
        dependencies as
        well as your binary, but the result will likely interfere with
        an installed
        MacPorts or Fink on the installer's machine.


    I link the *.a static libs into my executable and then test if it
    works.  Then I do a "mv /opt/local /opt/local_disable" and test
    again.  It works :)  Though only for Intel.  The PPC code crashes at
    some point (regardless of the presence of /opt/local).

Did you check all libraries/binaries for their architecture with lipo
-info <blahblah>?
GCC 4.2.1 is designed to build ppc64 for ppc architecture. I assume that
part of your libs/bins is ppc64 instead of ppc.

lipo -info says: "Non-fat file: Foo is architecture: ppc7400" These are the Qt frameworks (built on my system, with the default compiler Qt uses when you choose "32-bit PPC", which is GCC 4.2.1).

However, on the libraries build with MacPorts, lipo says ppc instead of ppc7400: "Architectures in the fat file: libvorbis.a are: i386 ppc"

_______________________________________________
macports-users mailing list
[email protected]
http://lists.macosforge.org/mailman/listinfo.cgi/macports-users

Reply via email to