libomp is provided by
- LLVM, which gets pulled in with FLANG
- gcc
So, we *must* only depend on devel/openmp when
- the OPENMP knob is on
- FLANG is off (guaranteed to be off for all archs except amd64)
- clang is the C compiler.
But, we cannot modify LIB_DEPENDS after CHOSEN_COMPILER_TYPE is defined, so
for now, limit the OPENMP knob for i386 and amd64 where clang is the
default compiler.
This will be moot after 12.1 when openmp will be in base.
PR: 240279
Reported by: pkubaj
Reviewed by: pkubaj
as defined in Mk/bsd.default-versions.mk which has moved from GCC 8.3
to GCC 9.1 under most circumstances now after revision 507371.
This includes ports
- with USE_GCC=yes or USE_GCC=any,
- with USES=fortran,
- using Mk/bsd.octave.mk which in turn features USES=fortran, and
- with USES=compiler specifying openmp, nestedfct, c11, c++0x, c++11-lang,
c++11-lib, c++14-lang, c++17-lang, or gcc-c++11-lib
plus, everything INDEX-11 shows with a dependency on lang/gcc9 now.
PR: 238330
- Expire after the last version without /usr/lib/libomp.so
- Drop SOVERSION for seamless transition (i.e., avoid conditionals)
PR: 236907
Approved by: bapt (maintainer)
Differential Revision: https://reviews.freebsd.org/D19767
'CHANGES IN R 3.5.3' from
https://cran.r-project.org/doc/manuals/r-devel/NEWS.html
- writeLines(readLines(fnam), fnam) now works as expected, thanks to Peter
Meissner's PR#17528
- setClassUnion() no longer warns, but uses message() for now, when
encountering "non local" subclasses of class members
- stopifnot(exprs = T) no longer fails
defined via Mk/bsd.default-versions.mk which has moved from GCC 7.4 t
GCC 8.2 under most circumstances.
This includes ports
- with USE_GCC=yes or USE_GCC=any,
- with USES=fortran,
- using Mk/bsd.octave.mk which in turn features USES=fortran, and
- with USES=compiler specifying openmp, nestedfct, c11, c++0x, c++11-lang,
c++11-lib, c++14-lang, c++17-lang, or gcc-c++11-lib
plus, as a double check, everything INDEX-11 showed depending on lang/gcc7.
PR: 231590
After a discussion on the mailing list on moving manpages to
${PREFIX}/share/man for consistency with base where it is
installed in usr/share/man, it appeared the same should happen
to GNU info files which were installed under share in base and
not in ports.
Now texinfo is not in base on any of the supported version of FreeBSD
it is possible to proceed to this move and it is easier to do than
the manpage change.
Other benefit than consistency are less patching: all build tools but
cmake are expecting info files to be under share/info and cmake (patched here)
was having an exception for BSD so the patch makes FreeBSD case less
specific for them
Bump revision of all impacted ports
PR: 232907
exp-run by: antoine
Differential Revision: https://reviews.freebsd.org/D17816
Simplify some ports where DragonFlyBSD no longer needs to be special-cased.
Submitted by: rene
Reviewed by: bapt, jbeich
Differential Revision: https://reviews.freebsd.org/D17724
in the ports tree (via Mk/bsd.default-versions.mk and lang/gcc) which
has now moved from GCC 6 to GCC 7 by default.
This includes ports
- featuring USE_GCC=yes or USE_GCC=any,
- featuring USES=fortran,
- using Mk/bsd.octave.mk which in turn features USES=fortran, and those
- with USES=compiler specifying one of openmp, nestedfct, c11, c++0x,
c++11-lib, c++11-lang, c++14-lang, c++17-lang, or gcc-c++11-lib.
PR: 222542
When ATLAS is built locally and then injected into poudriere (or other
similar shenanigans) it does not make sense to block building of R.
Approved by: jrm (maintainer)
Other port changes:
- Add armv6 and armv7 to ONLY_FOR_ARCHS; armv6 and armv7 were tested by mmel@
and working on 11.
- Remove LIBR option and always build the shared library. This was requested by
yuri@ and a RFC was posted on ports@ with no response.
- Exclude OPENMP option for everything except amd64 and i386.
- Turn on OPENMP option by default to match upstream.
- Turn off TCLTK option by default.
- LLVM, which gets pulled in with FLANG, provides libomp, so only depend on
devel/openmp, when GFORTRAN and OPENMP are on.
- Rename USES=fortran:gcc to the more appropriate USES=fortran:gfortran.
- Change the options in math/R and math/libRmath so the choice becomes
flang or default instead of flang or gfortran. Use default Fortran
compiler again on amd64.
Also
- Build Fortran code with flang by default on amd64.
- Remove LTO option. Upstream does not turn it on by default, it has
caused problems with each GCC upgrade, and any benefit is unclear.
- Do not specify USES=compiler:openmp since that just forces GCC.
(via Mk/bsd.default-versions.mk and lang/gcc) which has moved from
GCC 5.4 to GCC 6.4 under most circumstances.
This includes ports
- with USE_GCC=yes or USE_GCC=any,
- with USES=fortran,
- using Mk/bsd.octave.mk which in turn features USES=fortran, and
- with USES=compiler specifying openmp, nestedfct, c++11-lib, c++11-lang,
c++14-lang, c++0x, c11, or gcc-c++11-lib.
PR: 219275
Specifying TCLTK_USES=tk:84+ was pulling in version 8.4 for package builds on
cluster, however the Tcl/Tk 8.4 ports are being deprecated. Simply using
TCLTK_USES=tk pulls in the default version. [1]
While here, silence a portlint warning about the location of ONLY_FOR_ARCHS and
use DISTVESION instead of PORTVERSION.
Reported by: gahr [1]
permitting FLANG only with RBLAS
- Only apply SONAME patches when the relevant options are chosen. This
prevents install errors when LIBR is off.
- Only permit the FLANG option when RBLAS is chosen, because FLANG causes
some configure-time checks to fail and the shared RBLAS library is fallen
back upon, but not properly packaged, which causes runtime issues.
- Bump PORTREVISION
Reported by: rhurlin@gwdg.de
Differential Revision: https://reviews.freebsd.org/D11726
The configure script now detects the correct zlib version, so setting the
r_cv_header_zlib_h is no longer necessary.
Approved by: swills (mentor, implicit)
