The thread, which may have reached EOL, has lost it's way. I have been asking more about the general setup for cross-compiling more than about the GNU SDK. As I intend to deploy the Raspberry image as openSUSE, I am most interested in that environment. If I just want to test that the code compiles, I already compile for x86 openSUSE (12.3, 13.1, 42.2, Tumbleweed) and Windows (32- and 64-bit cross-compile with mingw). And as lots of the code is related to hardware access, testing will have to happen on a real device. The benefit of cross-compiling for ARM is outweighed by the chance of error maintaining an ARM install in parallel with the x86 install.
I have my answer: compile on a Raspberry. It has the source mounted via NFS and the results are available immediately for packaging on the build server. Our make system allows parallel concurrent distributed builds for multiple architectures. The Raspberry device will be just another one. If I want to get fancy, all I lack on the Raspberry is subversion. But since the sources are shared over NFS, that is not a real issue. I just check in any Raspberry related changes from the x86 host. Thanks for the discussion! -- Roger Oberholtzer -- To unsubscribe, e-mail: [email protected] To contact the owner, e-mail: [email protected]
