On 2018/11/15 22:51, Klemens Nanni wrote: > Proper diff incl. patch-configure changes, thanks rsadowski.
FWIW this builds OK on i386. One failure in "make test" (see below), no idea how important that is. It's not a good machine for real-world runtime tests. Watch out for programs which have SDL in WANTLIB and are relying on ffmpeg for their LIB_DEPENDS path to devel/sdl.. $ cat tests/data/fate/checkasm-aacpsdsp.err checkasm: using random seed 3325342763 SSE: - aacpsdsp.add_squares [OK] - aacpsdsp.mul_pair_single [OK] - aacpsdsp.hybrid_analysis [OK] - aacpsdsp.hybrid_analysis_ileave [OK] - aacpsdsp.hybrid_synthesis_deint [OK] SSE3: - aacpsdsp.add_squares [OK] - aacpsdsp.hybrid_analysis [OK] test failed comparing 977.599 with 977.604 (abs diff=0.00512695 with EPS=0.005) ps_stereo_interpolate_sse3 (aacpsdsp.c:206) - aacpsdsp.stereo_interpolate [FAILED] SSE4.1: - aacpsdsp.hybrid_synthesis_deint [OK] checkasm: 1 of 10 tests have failed