sizeof(GstRegistry) is 144, expected 72 sizeof(GstRegistryClass) is 288, expected 144 sizeof(GstSegment) is 104, expected 88 sizeof(GstStructure) is 40, expected 20 sizeof(GstSystemClock) is 288, expected 200 sizeof(GstSystemClockClass) is 352, expected 176 sizeof(GstTagSetterIFace) is 16, expected 8 sizeof(GstTask) is 160, expected 80 sizeof(GstTaskClass) is 280, expected 140 sizeof(GstTrace) is 32, expected 20 sizeof(GstAllocTrace) is 24, expected 16 sizeof(GstTypeFind) is 64, expected 32 sizeof(GstTypeFindFactory) is 216, expected 108 sizeof(GstTypeFindFactoryClass) is 304, expected 152 sizeof(GstURIHandlerInterface) is 88, expected 44 sizeof(GstValueTable) is 64, expected 32 sizeof(GstXML) is 128, expected 64 sizeof(GstXMLClass) is 288, expected 144 0%: Checks: 1, Failures: 1, Errors: 0 gstcheck.c:283:F:size check:tf:0: failed ABI check FAIL: gst/gstabi Running suite(s): GstBuffer tests/check/gst/struct_ppc32.h and tests/check/gst/struct_ppc64.h are almost the same, but I think on ppc64 pointers are bigger. In fact, copying tests/check/gst/struct_x86_64.h to tests/check/gst/struct_ppc64.h makes this test pass. So it seems the tests aren't correct, but better have the opinion of someone more experienced with ppc64 than me ;)
the same applies to tests/check/libs/struct_ppc64.h
filled upstream too
good work! gstreamer tests always failed, but gstreamer itself worked otherwise. I think we should create a patch. I'll do it once I'm next to my ppc64 machine again.
(In reply to comment #3) > good work! gstreamer tests always failed, but gstreamer itself worked > otherwise. > > I think we should create a patch. I'll do it once I'm next to my ppc64 machine > again. > I've made one, its attached to the upstream bugreport and they just applied it. I dunno if gstreamer herd wants to apply it or just wait for the next release.
thanks, alexis! I'd say let's wait for upstream releasing a new version.
media-libs/gstreamer-0.10.16 now passes tests but gst-plugins-base-0.10.16 fails :-( will investigate this one