Martijn van Beurden wrote: > > op 29-06-14 15:31, Erik de Castro Lopo schreef: > > I would appreciate it if people could kick this around, test on other > > platforms/architectures and test for performance regressions. > > These patches indeed fix the -Wcast-align warnings, but the new > MD5 tests fail > > > +++ libFLAC unit test: md5 > > > > testing FLAC__MD5Init ... OK > > testing that FLAC__MD5Final clears the MD5Context ... OK > > testing FLAC__MD5Accumulate (channels=1, bytes_per_sample=1) ... > > FAILED, expected MD5 sum b2bb8775b7d5bf59c36c8637293a4602 but > > got 419123315a22592e3170b67d0aa26916 > > However, using the binary on actual FLAC files (with flac -t) > doesn't give any errors of mismatching md5 sums, so maybe the > problem is in the test itself?
Possibly. I will investigate. Fortunately, I got the same error in my QEMU armhf chroot, but I didn't trust QEMU. > ARM is big endian IIRC, maybe that could be the cause? ARM is actully bi-endian. The endian-ness is often configured by an external pin on the chip but some systems can swap endian-nesses at run time by updating a register but obviously this is not something we need to support. We just need to support whatever endian-ness is detected at configure time. Erik -- ---------------------------------------------------------------------- Erik de Castro Lopo http://www.mega-nerd.com/ _______________________________________________ flac-dev mailing list flac-dev@xiph.org http://lists.xiph.org/mailman/listinfo/flac-dev