This only affects "Created by" lines with one exception: devel/uclcmd. There the maintainer is changed. This was overlooked in r416918.
Approved by: junovitch (mentor)
Add patch for integer overflow in zipimport module to all our python ports.
While I'm here, get rid of -f flag in ${RM} invocation, because ${RM} already
expands to rm -f, so in result we are getting something like:
/bin/rm -f -f /wrkdirs/usr/ports/lang/python35/work/stage/usr/local/lib/libpython3.so
PR: 210325
Submitted by: Vladimir Krstulja <vlad-fbsd@acheronmedia.com>
Security: 1d0f6852-33d8-11e6-a671-60a44ce6887b
With hat: python
--- loadpath.o ---
loadpath.c:30:2: error: RUBY_EXEC_PREFIX must be defined
^
loadpath.c:41:50: error: use of undeclared identifier 'RUBY_LIB_VERSION'
typedef char ruby_lib_version_string[(int)sizeof(RUBY_LIB_VERSION) - 2];
Sponsored by: Absolight
- Add RUN_DEPENDS on devel/pcc-libs port on the same version
- Adapt PLIST_FILES
- Add THREADS option to "Enable Thread-local storage (TLS)"
- Add new patch for ${WRKSRC}/cc/cc/cc.c file, related to inverted
MULTIARCH_PATH define check
- Replace some static patches with sed patches
- Remove unneeded patches for this version
PR: 196642
Submitted by: Sender Ghost <lightside@gmx.com>
Approved by: joel (maintainer)
By Makefile restriction, gcc6-aux can only be built on x86-64 and i386.
This restriction is due to a lack of a bootstrap compiler required for
Ada. Thus, the recently added ARMv6 patch has no effect (plus gcc6-aux
is updated by diff-* patchsets and single patches can interfer with
that scheme.)
The default mode for C++ is now -std=gnu++14 instead of -std=gnu++98.
Type-based alias analysis now disambiguates accesses to different pointers.
This improves precision of the alias oracle by about 20-30% on higher-level
C++ programs. Programs doing invalid type punning of pointer types may now
need -fno-strict-aliasing to work correctly.
Value range propagation now assumes that the this pointer of C++ member
functions is non-null. This eliminates common null pointer checks but also
breaks some non-conforming code-bases (such as Qt-5, Chromium, KDevelop).
As a temporary work-around -fno-delete-null-pointer-checks can be used.
Wrong code can be identified by using -fsanitize=undefined.
There have been significant improvements around link-time optimization
and inter-procedural optimization and some support for OpenACC (though
probably not tested on FreeBSD) and version 4.5 of the OpenMP specification.
Source locations for the C and C++ compilers are now tracked as ranges
and diagnostics can now contain "fix-it hints".
New command-line options include -Wshift-negative-value, -Wshift-overflow,
-Wtautological-compare, -Wnull-dereference, -Wduplicated-cond, and
-Wmisleading-indentation.
C++ Concepts are now supported when compiling with -fconcepts, and there
are several other improvements around support for newer C++ features as
well as in the libstdc++ library.
Fortran now features Fortran 2008 SUBMODULE support, Fortran 2015 EVENT_TYPE,
EVENT_POST, EVENT_WAIT, and EVENT_QUERY support, and improved support for
OpenMP and OpenACC.
A lot has happened on the AArch64 and ARM fronts, on x86-64 there is now
Skylake with AVX-512 support, AMD instructions monitorx and mwaitx, and
support for address spaces __seg_fs, __seg_gs, and __seg_tl, as well as
AMD Zen (family 17h) processors, and basic support has been added for POWER9.
https://gcc.gnu.org/gcc-6/changes.html has a more extensive set of
changes and https://gcc.gnu.org/gcc-6/porting_to.html has a solid
overview of issue you may encountering porting to this new version.
The numbering scheme changed to presumably match years (e.g. 17 => 2017)
Bump everything with a build dependency as there is a matching soname
change. Everything still builds on a DF exp-run.
