math/reduce: fix the build with LLVM 11
reduce started to fail with the default switch to -fno-common in LLVM 11
(and GCC 10). This patch was written then confirmed to be the fix that
upstream also provided in late March.
Approved by: ports-secteam (implicit, -fno-common build fix)
math/units: Add a default option READLINE to tuggle readline support
This makes sure the readline support is tuggled in the build time and not
affected by the libreadline installed or not during build/run time.
PR: 248631
Submitted by: ports-units@shalott.net
Approved by: jharris@widomaker.com (maintainer)
Approved by: portmgr (blanket: build, run, dependency fix)
math/cadical: fix build on GCC architectures
Use | instead of , in the sed command to workaround for -Wl,rpath in CXXFLAGS.
Approved by: portmgr (fix build blanket)
graphics/ilmbase, graphics/openexr: security update to v2.5.3
ChangeLog:
https://github.com/AcademySoftwareFoundation/openexr/releases/tag/v2.5.3
"Patch release with various bug/security fixes and build/install fixes, plus a performance optimization:
* Various sanitizer/fuzz-identified issues related to handling of invalid input
* Fixes to misc compiler warnings [...]
* Read performance optimization (#782)
* Fix for building on non-glibc (#798)
* Fixes to tests"
The full name of shared libraries changed from *.25.0.1 to *.25.0.2
=> PORTREVISION bumps for all ports that use either ilmbase or openexr
coming up in separate commit to ease MFH.
All test builds of direct users succeeded on 12.1 amd64.
Bump PORTREVISION of all ports using ilmbase or openexr.
Security: b1d6b383-dd51-11ea-a688-7b12871ef3ad
Approved by: ports-secteam@ (joneum@)
math/wxmaxima: fix build on GCC architectires
Use C++11 compiler:
CMake Error in src/CMakeLists.txt:
Target "wxmaxima" requires the language dialect "CXX11" (with compiler
extensions), but CMake does not know the compile flags to use to enable it.
Approved by: portmgr (fix build blanket)
math/cocoalib: fix build on GCC architectures
Use C++11 compiler:
/wrkdirs/usr/ports/math/cocoalib/work/CoCoALib-0.99710/include/CoCoA/TmpPBMill.H:263: error: 'nullptr' was not declared in this scope
Approved by: portmgr (fix build blanket)
math/clasp: fix build on GCC architectures
Use C++11 compiler:
CMake Error in app/CMakeLists.txt:
Target "clasp" requires the language dialect "CXX11" (with compiler
extensions), but CMake does not know the compile flags to use to enable it.
Approved by: portmgr (fix build blanket)
- remove CONFLICTS with lang/python3 (DP)
- remove future CONFLICTS with blas/lapack (PR 247542)
- add a pkg-message for SAGE_ROOT / SAGE_LOCAL (DP)
- remove the check of R´s Makevars.site (DP)
see <https://trac.sagemath.org/ticket/29943>
- use math/e-antic from the ports tree
- make BROKEN on -CURRENT untill a solution is found / Python
Reported by: Dima Pasechnik <dimpase (at) gmail.com> (DP)
which should never be set on the FreeBSD package building cluster
whatsoever:
g++9: error: unrecognized command line option '-march=native'; did you mean '-mcpu=native'?
Unfortunately this is necessary but not sufficient for powerpc64:
/wrkdirs/usr/ports/math/ogdf/work/OGDF/src/ogdf/basic/System.cpp:78:11: fatal error: cpuid.h: No such file or directory
Approved by: portmgr (tier-2 blanket)
- The build is purely CMake-based now and doesn't use any shell
or Python helper scripts
- Keep PLIST_SUB+=COIN="" to reduce the churn on the pkg-plist,
will be cleaned up separately (COIN support is now mandatory)
- Chase 301 Moved Permanently redirect in the WWW line of the
port description while here
giacxas expects mallinfo in malloc.h
On systems before 13 the detection of malloc.h fails, but on 13 it succeeds but compilation then fails due to a missing mallinfo function.
Reported by: thierry, fallout
This version fixes a warning when building with a strict warnings level
and significantly speeds up ceil() library function.
Approved by: antoine (implicit)
This port won't build with base GCC, use GCC from ports.
Also add -mfloat128, which requires -mvsx. This will mean the binary will need to be used on POWER7 or newer, because -mfloat128 uses VSX.
