Date: Wed, 12 Feb 2020 01:01:25 +0100 (CET) From: Gerald Pfeifer <gerald@pfeifer.com> To: Piotr Kubaj <pkubaj@FreeBSD.org>, powerpc@FreeBSD.org Cc: ports-committers@freebsd.org, svn-ports-all@freebsd.org, svn-ports-head@freebsd.org Subject: Re: svn commit: r525704 - head/math/hmat-oss Message-ID: <alpine.LSU.2.21.2002120045231.5016@anthias.pfeifer.com> In-Reply-To: <202002101323.01ADNuln021504@repo.freebsd.org> References: <202002101323.01ADNuln021504@repo.freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
[ adding powerpc@ ]
On Mon, 10 Feb 2020, Piotr Kubaj wrote:
> Log:
> math/hmat-oss: fix build on GCC architectures
>
> After OpenBLAS update, new GCC is required:
> CMake Error at /usr/local/share/cmake/Modules/FindPackageHandleStandardArgs.cmake:146 (message):
> Could NOT find CBLAS (missing: CBLAS_LIBRARIES)
:
> +.if ${CHOSEN_COMPILER_TYPE} == gcc
> +USE_GCC= yes
> +.endif
This made me realize something bizarre: Mk/Uses/compiler.mk has
.if ${_COMPILER_ARGS:Mopenmp}
.if ${COMPILER_TYPE} == clang
USE_GCC= yes
CHOSEN_COMPILER_TYPE= gcc
.endif
.endif
If I read this correctly, on platforms with GCC as the system compiler
this means the rotten GCC 4.2 is used for something new and evolving
like OpenMP?
Really, OpenMP should imply USE_GCC=yes instead of even trying to
rely on GCC 4.2. That would save you the effort of having to make
such changes one by one and also simplify the ports tree.
*And* it would align powerpc* more closely with i386/amd64 by using
the same compiler for such ports.
I'll be happy to prepare a patch for Mk/Uses/compiler.mk, but we'd
need you (and/or other powerpc* folks) to give it a try and adjust
things if necessary. Would you like to?
Gerald
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?alpine.LSU.2.21.2002120045231.5016>