Date: Sat, 11 Jan 2020 04:11:34 +0000 From: bugzilla-noreply@freebsd.org To: ports-bugs@FreeBSD.org Subject: [Bug 238773] multimedia/x265: Only highest bit-depth profile is built when multiple (bit-depth OPTIONS) are selected Message-ID: <bug-238773-7788-KCNoeEj55H@https.bugs.freebsd.org/bugzilla/> In-Reply-To: <bug-238773-7788@https.bugs.freebsd.org/bugzilla/> References: <bug-238773-7788@https.bugs.freebsd.org/bugzilla/>
next in thread | previous in thread | raw e-mail | index | archive | help
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=3D238773 --- Comment #18 from Jamie Landeg-Jones <jamie@catflap.org> --- (In reply to Mikhail Teterin from comment #15) Hi Mikhail! The "promised behavior" was never correct. It implied 8 bit build, along wi= th optional 10bit and 12bit. It never stated what the default bitdepth would be if not specified on the command line. As it stood, only the highest bitdepth selected was actually build, and hen= ce, it became the defacto default. Daniels patch does indeed build all bitdepths into one library. The default bitdepth will now be 8bit, with 10 and 12 bit optional on the x265 command line. I agree with Daniel that this is the best solution - if you want a higher b= it depth, it's easily selectable. (I think the bitdepth should be specified regardless anyway..) To emulate the previous behaviour you'd need to have a compile time option = to configure it at build time within the ports Makefile. I think this is overk= ill, and will make support/pkg harder to manage. I did suggest mentioning that this may make an unexpected change for some u= se cases, but as Daniel pointed out, it's clear at use time what bit depth is being used. I'm not sure where the extra libraries and include files comes from - Danie= ls patch simply bundles the whole lot into one (ie. the plist doesn't change) As for the output, use "x265 --version". This is what I get with Daniels patch applied: % x265 --version x265 [info]: HEVC encoder version 3.2.1+1-b5c86a64bbbe x265 [info]: build info [Unk-OS][clang 7.0.1][64 bit] 8bit+10bit+12bit x265 [info]: using cpu capabilities: MMX2 SSE2Fast LZCNT SSSE3 SSE4.2 AVX F= MA3 BMI2 AVX2 , and a quick test encode of all three profiles worked, producing different binaries, that ffprobe showed had the correct bit depth selected. As for the architecture issues, I must admit I did noting other than see if= it compiled on my box, which is Freebsd12 with the following flags: 04:09 (7) "/tmp" jamie@thompson% show-clang-native-features Clang "native" target for this machine: skylake (Triple: x86_64-unknown-freebsd12.0) Features enabled (+) : sse2, cx16, sahf, prfchw, bmi2, fsgsbase, xsavec, popcnt, aes, xsaves, mmx, rdseed, clflushopt, xsave, invpcid, avx, fma, bmi, rdrnd, sse4.1, sse4.2, avx2, sse, lzcnt, pclmul, f16c, ssse3, sgx, cmov, mo= vbe, xsaveopt, adx, sse3 Features disabled (-) : tbm, avx512ifma, sha, gfni, fma4, vpclmulqdq, cldem= ote, ptwrite, avx512bitalg, movdiri, avx512er, avx512vnni, avx512vpopcntdq, pcon= fig, clwb, avx512f, clzero, pku, lwp, rdpid, xop, waitpkg, movdir64b, sse4a, avx512bw, avx512vbmi2, avx512vl, avx512cd, vaes, rtm, mwaitx, wbnoinvd, prefetchwt1, shstk, avx512vbmi, avx512dq, avx512pf 04:09 (8) "/tmp" jamie@thompson% cc + cc -Wall -std=3Dgnu99 -O2 -pipe -march=3Dnative -mtune=3Dnative -ftree-ve= ctorize -fstack-protector-strong Hope this helps! cheers! --=20 You are receiving this mail because: You are the assignee for the bug.=
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-238773-7788-KCNoeEj55H>