On 06/11/18 11:23, Dr. Thomas Orgis wrote:
Am Mon, 11 Jun 2018 11:15:15 -0500 schrieb Jason Bacon <outpaddling%yahoo.com@localhost>:I would be inclined to cooperate with upstream's choice of soversion, but this does present a potential problem of an soversion going backwards with the next update. I'm not sure what the ramifications are for pkgsrc...How is the process to come to a decision there? I must admit that I am not really affected by the update procedure as I always build a fresh pkgsrc tree and keep it that way. IMHO, a switch in the version and an additional symlink libblas.so.4 → libblas.so.3 should work to keep binaries happy. I'll investigate how the cmake build might work for us … not sure yet if all switches are there. Alrigthy then, Thomas
Here's the sum total of committed packages using blas and lapack: [root@centosdev pkg-dev]# pkg-grep blas/build -l|grep -v ^wip biology/mpqc/Makefile biology/plink/Makefile lang/lush/Makefile math/harminv/Makefile math/ipopt/Makefile math/itpp/Makefile math/lapack/Makefile math/octave/Makefile math/py-numpy/Makefile math/py-scikit-learn/Makefile math/R-gstat/Makefile math/R/Makefile math/R-quantreg/Makefile math/R-RandomFields/Makefile math/R-wle/Makefile [root@centosdev pkg-dev]# pkg-grep lapack/build -l|grep -v ^wip biology/mpqc/Makefile biology/plink/Makefile lang/lush/Makefile math/harminv/Makefile math/ipopt/Makefile math/itpp/Makefile math/octave/Makefile math/py-numpy/Makefile math/py-scikit-learn/Makefile math/R/Makefile math/R-quantreg/MakefileShouldn't be too hard to determine the impact of an soversion change. Since they all just include the bl3 file, should be a pretty easy fix too. The .4 symlink would probably do it, if it's even necessary.