These ports link some non-PIC code, which fails with lld as it defaults
to disallowing relocations against read-only segments. For i386 we can
just add -znotext unconditionally: for GNU BFD ld it just affirms BFD's
existing default.
PR: 214864
Approved by: portmgr (lld blanket)
Sponsored by: The FreeBSD Foundation
This resolves a ninja-related issues, corrects passing private args to linker
while building libruby itself (without external LDFLAGS), exports the full
symbols table, and pet 'pkg check -d'. Issues with weechat, vim, facter,
rubygem-hpricot and rubygem-rdiscount should be fixed. See [1] for long story
of hunting and fixing these issues.
Bump PORTREVISIONS accordingly.
PR: 230238 [1], 231361
Submitted by: fuffy
Approved by: fuffy
mv: rename /wrkdirs/usr/ports/lang/rust-nightly/work/rustc-nightly-src/src/vendor/backtrace-sys/src/libbacktrace/config/libtool.m4 to /wrkdirs/usr/ports/lang/rust-nightly/work/rustc-nightly-src/src/vendor/backtrace-sys/src/libbacktrace/config/libtool.m4.dont-fix: No such file or directory
*** Error code 1
As the port appears to be a staging ground for future updates don't
deprecate yet.
PR: 229826
Reported by: pkg-fallout
- libgit2 in ports is nowadays newer than bundled version
- libssh2 was already unbundled until 1.29.0 update
- Unbundle libssh2 for USES=cargo (unused atm)
PR: 231411
Approved by: rust (dumbbell, tobik)
It was initially got from ruby23, but recent ruby releases have some
differences in runtime linking, which was not hit by exp-run, and caused
trouble with some ports (e.g. devel/leatherman, sysutils/facter).
PR: 230238
Submitted by: fluffy
Reported by: freebsd@ptty.com
Approved by: fluffy
This port links some non-PIC code, which fails with lld as it defaults
to disallowing relocations against read-only segments. For i386 we can
just add -znotext unconditionally: for GNU BFD ld it just affirms BFD's
existing default.
PR: 214864, 230207
Approved by: bapt
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D17193
Example error:
/usr/bin/ld: error: can't create dynamic relocation R_386_32 against
symbol: caml_last_return_address in readonly segment; recompile
object files with -fPIC
This port links some non-PIC code, which fails with lld as it defaults
to disallowing relocations against read-only segments. For i386 we can
just add -znotext unconditionally: for GNU BFD ld it just affirms BFD's
existing default.
PR: 214864, 230209
Approved by: bapt
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D17200
Release notes:
* https://blog.rust-lang.org/2018/07/20/Rust-1.27.2.html
* https://blog.rust-lang.org/2018/08/02/Rust-1.28.html
* https://blog.rust-lang.org/2018/09/13/Rust-1.29.html
Up to and including Rust 1.27.x, the Rust build system shelled out to
a configure script to detect the presence and usability of libunwind.
Since Rust 1.28.0, it's using a static result in a `build.rs` file and
expects libunwind to be used. It was not the case on FreeBSD so far, so
we need a patch to this `build.rs` to disable that. We still need to
study if the FreeBSD port should use libunwind and what to do with this
patch. But this problem prevented the update to Rust 1.28.0 already, so
enough delay.
The update also comes with a patch to a few `USE_GECKO`-based ports such
as Firefox [1]. Their configure script has some asumptions on the output
of `rustc --print target-list` which are not true anymore. The patch was
already committed upstream.
The aarch64 version is still marked as BROKEN because I didn't find the
time to work on it. As a consequence, there is also no aarch64 bootstrap
for Rust 1.29.0.
PR: 229826
Approved by: jbeich [1]
Obtained from: https://bugzilla.mozilla.org/show_bug.cgi?id=1479540 [1]
Differential Revision: https://reviews.freebsd.org/D17178
This change merges the following upstream pull request into the Ruby
interpreter ports:
https://github.com/ruby/ruby/pull/1954
Adding this patch to these ports will ensure that once we patch up
qsort_r() to be compatible with glibc, Ruby will automatically pick them
up. Ruby should also build fine without this patch, but this will cause
it to use its own implementation, which blows up the binary size
slightly.
Poudriere runs seem to pass for 11.x amd64. Logs indicate that this
doesn't negatively affect the existing qsort_r() detection:
checking whether qsort_r is GNU version... no
checking whether qsort_r is BSD version... yes
PR: 231256
Approved by: sunpoet
Differential Revision: https://reviews.freebsd.org/D17157
Change x11/xorgproto to become a build time dependency when added to
USE_XORG. Change the dependency to be on the port, rather than a file the
port installs.
Fix fallout.
Bump portrevision on depending ports.
PR: 230909
Reviewed by: eadler
Approved by: portmgr (antoine)
Obtained from: https://github.com/FreeBSDDesktop/freebsd-ports/tree/feature/xorgproto
exp-run: antoine
Differential Revision: https://reviews.freebsd.org/D16906
20180821:
- In fil_dos.c in the function setupFiles the handling of ctrl-c has
been improved such that ctrl-c is sent immediate (without the need to
press return afterwards). Additionally redirected stdin, stdout and
stderr files are switched to binary mode now.
- In the compiler (in comp/const.s7i) the actions BIN_XOR, ENU_ICONV2,
ENU_ORD2, INT_ICONV1 and INT_ICONV3 have been added to the list of
special actions.
- The batch files bin/call_gcc.bat and bin/call_ar.bat have been
improved to set environment variables with the command 'set'. This
is necessary to execute them successfully under Wine.
- The description of log, log10 and log2 in the manual has been
improved.
- In flt_rtl.c the functions fltLog, fltLog10 and fltLog2 have been
added. This functions are based on the original functions log(),
log10() and log2() and work according to IEEE 754.
- The configuration values LOG_FUNCTION_OKAY, LOG10_FUNCTION_OKAY and
LOG2_FUNCTION_OKAY have been added to cc_conf.s7i.
- The configuration value HAS_LOG2 has been removed from cc_conf.s7i.
- The compiler (s7c) has been improved to use the functions fltLog,
fltLog10 and fltLog2.
- In comp/inline.s7i the functions push_inline_func_param,
pop_inline_func_param, push_inline_proc_param and
pop_inline_proc_param have been improved to support inline functions
with 'attr' parameters.
- The example program diff7.sd7 has been improved.
- Documentation comments have been improved in math.s7i and tar.s7i.
- In chkflt.sd7 the tests for the operators 'parse' and 'sci' and for
the function str() have been improved.
- The program chkccomp.c has been improved to define the macros
POW_OF_NEGATIVE_OKAY, LOG_OF_NAN_OKAY, LOG_OF_ZERO_OKAY,
LOG_OF_NEGATIVE_OKAY, LOG10_OF_NAN_OKAY, LOG10_OF_ZERO_OKAY,
LOG10_OF_NEGATIVE_OKAY, LOG2_OF_NAN_OKAY, LOG2_OF_ZERO_OKAY,
LOG2_OF_NEGATIVE_OKAY, MEMCMP_RETURNS_SIGNUM and
WMEMCMP_RETURNS_SIGNUM.
- In chkccomp.c the function determineFseekFunctions has been improved
to make tests that actually use fseek() and ftell().
- Calls of logFunction have been added in cmd_rtl.c, str_rtl.c,
strlib.c and striutl.c.
- In bstlib.c the function bst_cmp() has been improved to be based on
bstCmp().
- The functions strCompare(), str_cmp() and bstCmp() have been
improved to use different code depending on MEMCMP_RETURNS_SIGNUM
and WMEMCMP_RETURNS_SIGNUM.
- In s7c.sd7 the function canTakeAddress has been improved to return
FALSE for a CONSTENUMOBJECT.
Disable the build/use of libssp/gets-chk since FreeBSD 12 and later
do not feature gets() any longer.
(I was planning to make this part of a routine update, alas last week's
snapshot was broken on i386-unknown-freebsd10.x and so will this week's
still.)
PR: 222796, 231066
Differential Revision: https://reviews.freebsd.org/D12298
I was too optimistic about the fix from r478932 and it quickly falls
to pieces with a higher number (> 16) of jobs. Upstream has never
supported building in parallel either.
Disable the build/use of libssp/gets-chk since FreeBSD 12 and later
do not feature gets() any longer.
PR: 222796, 231066
Differential Revision: https://reviews.freebsd.org/D12298
The last upstream release supporting KDE4 was 10 months ago,
https://www.kde.org/announcements/announce-applications-17.08.3.php ,
and the kde@ team for FreeBSD has supported modern KDE (i.e. KDE
Frameworks, Plasma 5 and KDE Applications) for 4 months. Give a 4
month deprecation period to allow users to switch over.
Also deprecated: multimedia/kmplayer-kde4 (permission from jhale@)
Not deprecated, need to contact maintainers:
misc/krecipes-kde4
net-im/kmess-kde4
net/smb4k-kde4
textproc/uim-kde4
- Bump PORTREVISION for package change
Support for building --without-threads has been removed.
The threading module is now always available.
PR: 230605
Submitted by: Roland Smith <rsmith@xs4all.nl>
Reference: https://docs.python.org/3/whatsnew/3.7.html
CHICKEN is a compiler for the Scheme programming language. CHICKEN
produces portable, efficient C, supports almost all of the R5RS
Scheme language standard, and includes many enhancements and
extensions.
WWW: http://www.call-cc.org/
CHICKEN 4 and 5 are not fully compatible. Import the upcoming
CHICKEN 5.0.0 as a new port. Binaries and libraries are suffixed
with a 5 to avoid conflicts with lang/chicken.
Changes: https://code.call-cc.org/dev-snapshots/2018/08/11/NEWS
PR: 230535
2018-08-20 net/openmq: Broken for more than 5 months
2018-08-20 net/coda6_server: Broken for more than 5 months
2018-08-20 net/ceph-devel: Broken for more than 5 months
2018-08-20 java/jgrapht: Broken for more than 5 months
2018-08-20 mail/mutt14: mutt 1.4.x has been out of support for 10+ years use mail/mutt
2018-08-20 www/pydio: Broken for more than 5 months
2018-08-20 security/fastd-devel: Broken for more than 5 months
2018-08-20 security/fastd: Broken for more than 5 months
2018-08-20 devel/bhyve-vm-goagent: Broken for more than 5 months
2018-08-20 devel/cherivis-devel: Depends on broken and expiring cheritrace-devel
2018-08-20 devel/cheritrace-devel: Broken for more than 5 months
2018-08-20 devel/javolution: Broken for more than 5 months
2018-08-20 devel/bisoncpp: Depends on broken and expiring libbobcat
2018-08-20 databases/php5-pdo_cassandra: Broken for more than 5 months
2018-08-20 games/crafty-open-large: Broken for more than 5 months
2018-08-20 games/crafty-open-enormous: Broken for more than 5 months
2018-08-20 games/crafty-open-medium: Broken for more than 5 months
2018-08-20 math/jakarta-commons-math: Broken for more than 5 months
2018-08-20 math/hfst: Broken for more than 5 months
2018-08-20 math/octave-forge-communications: Broken for more than 5 months
2018-08-20 lang/v8-devel: Broken for more than 5 months
2018-08-20 sysutils/rubygem-fluentd010: Broken for more than 5 months
backend/src/backend/program.cpp:797:34: error: reference to type 'const llvm::Module' could not bind to an lvalue of type 'llvm::Module *'
llvm::WriteBitcodeToFile(*out_module, ostream);
^~~~~~~~~~~
/usr/local/llvm70/include/llvm/Bitcode/BitcodeWriter.h:129:41: note: passing argument to parameter 'M' here
void WriteBitcodeToFile(const Module &M, raw_ostream &Out,
^
backend/src/llvm/llvm_bitcode_link.cpp:343:36: error: no matching function for call to 'CloneModule'
llvm::Module * linked_module = llvm::CloneModule((llvm::Module*)mod).release();
^~~~~~~~~~~~~~~~~
/usr/local/llvm70/include/llvm/Transforms/Utils/Cloning.h:52:25: note: candidate function not viable: no known conversion from 'llvm::Module *' to 'const llvm::Module' for 1st argument; dereference the argument with *
std::unique_ptr<Module> CloneModule(const Module &M);
^
/usr/local/llvm70/include/llvm/Transforms/Utils/Cloning.h:53:25: note: candidate function not viable: requires 2 arguments, but 1 was provided
std::unique_ptr<Module> CloneModule(const Module &M, ValueToValueMapTy &VMap);
^
/usr/local/llvm70/include/llvm/Transforms/Utils/Cloning.h:60:1: note: candidate function not viable: requires 3 arguments, but 1 was provided
CloneModule(const Module &M, ValueToValueMapTy &VMap,
^
backend/src/llvm/ExpandLargeIntegers.cpp:230:5: error: use of undeclared identifier 'DEBUG'
DEBUG(dbgs() << "\tRecording as forward PHI\n");
^
backend/src/llvm/ExpandLargeIntegers.cpp:237:5: error: use of undeclared identifier 'DEBUG'
DEBUG(dbgs() << "\tTo: " << *To.Lo << "\n");
^
backend/src/llvm/ExpandLargeIntegers.cpp:238:5: error: use of undeclared identifier 'DEBUG'
DEBUG(dbgs() << "\tAnd: " << *To.Hi << "\n");
^
backend/src/llvm/ExpandLargeIntegers.cpp:247:5: error: use of undeclared identifier 'DEBUG'
DEBUG(dbgs() << "\tTo: " << *To << "\n");
^
backend/src/llvm/ExpandLargeIntegers.cpp:256:11: error: expected expression
DEBUG(if (!ForwardPHIs.empty()) dbgs() << "Patching forward PHIs:\n");
^
backend/src/llvm/ExpandLargeIntegers.cpp:261:7: error: use of undeclared identifier 'DEBUG'
DEBUG(dbgs() << "\t" << *F.Lo << "\n\t" << *F.Hi << "\n");
^
backend/src/llvm/ExpandLargeIntegers.cpp:389:3: error: use of undeclared identifier 'DEBUG'
DEBUG(dbgs() << "Expanding Large Integer: " << *Inst << "\n");
^
backend/src/llvm/llvm_to_gen.cpp:142:13: error: use of undeclared identifier 'createInstructionCombiningPass'; did you mean 'createFunctionInliningPass'?
MPM.add(createInstructionCombiningPass());// Clean up after IPCP & DAE
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
backend/src/llvm/llvm_to_gen.cpp:164:13: error: use of undeclared identifier 'createInstructionCombiningPass'; did you mean 'createFunctionInliningPass'?
MPM.add(createInstructionCombiningPass()); // Combine silly seq's
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
backend/src/llvm/llvm_to_gen.cpp:172:13: error: use of undeclared identifier 'createInstructionCombiningPass'; did you mean 'createFunctionInliningPass'?
MPM.add(createInstructionCombiningPass());
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
backend/src/llvm/llvm_to_gen.cpp:209:13: error: use of undeclared identifier 'createInstructionCombiningPass'; did you mean 'createFunctionInliningPass'?
MPM.add(createInstructionCombiningPass());
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
backend/src/llvm/llvm_to_gen.cpp:215:13: error: use of undeclared identifier 'createInstructionCombiningPass'; did you mean 'createFunctionInliningPass'?
MPM.add(createInstructionCombiningPass()); // Clean up after everything.
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
backend/src/llvm/llvm_to_gen.cpp:366:16: error: use of undeclared identifier 'createPromoteMemoryToRegisterPass'
passes.add(createPromoteMemoryToRegisterPass());
^
backend/src/llvm/llvm_to_gen.cpp:373:16: error: use of undeclared identifier 'createInstructionCombiningPass'; did you mean 'createFunctionInliningPass'?
passes.add(createInstructionCombiningPass()); // legalize will generate some silly instructions
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
backend/src/llvm/llvm_to_gen.cpp:380:16: error: use of undeclared identifier 'createLowerSwitchPass'; did you mean 'createLoopUnswitchPass'?
passes.add(createLowerSwitchPass()); // simplify cfg will generate switch-case instruction
^~~~~~~~~~~~~~~~~~~~~
backend/src/backend/gen_program.cpp:452:32: error: reference to type 'const llvm::Module' could not bind to an rvalue of type 'llvm::Module *'
llvm::WriteBitcodeToFile((llvm::Module*)prog->module, OS);
^~~~~~~~~~~~~~~~~~~~~~~~~~~
/usr/local/llvm70/include/llvm/Bitcode/BitcodeWriter.h:129:41: note: passing argument to parameter 'M' here
void WriteBitcodeToFile(const Module &M, raw_ostream &Out,
^
backend/src/backend/gen_program.cpp:543:29: error: no matching function for call to 'CloneModule'
llvm::Module* clone = llvm::CloneModule(src).release();
^~~~~~~~~~~~~~~~~
/usr/local/llvm70/include/llvm/Transforms/Utils/Cloning.h:52:25: note: candidate function not viable: no known conversion from 'llvm::Module *' to 'const llvm::Module' for 1st argument; dereference the argument with *
std::unique_ptr<Module> CloneModule(const Module &M);
^
/usr/local/llvm70/include/llvm/Transforms/Utils/Cloning.h:53:25: note: candidate function not viable: requires 2 arguments, but 1 was provided
std::unique_ptr<Module> CloneModule(const Module &M, ValueToValueMapTy &VMap);
^
/usr/local/llvm70/include/llvm/Transforms/Utils/Cloning.h:60:1: note: candidate function not viable: requires 3 arguments, but 1 was provided
CloneModule(const Module &M, ValueToValueMapTy &VMap,
^
backend/src/backend/gen_program.cpp:452:32: error: reference to type 'const llvm::Module' could not bind to an rvalue of type 'llvm::Module *'
llvm::WriteBitcodeToFile((llvm::Module*)prog->module, OS);
^~~~~~~~~~~~~~~~~~~~~~~~~~~
/usr/local/llvm70/include/llvm/Bitcode/BitcodeWriter.h:129:41: note: passing argument to parameter 'M' here
void WriteBitcodeToFile(const Module &M, raw_ostream &Out,
^
Obtained from: upstream (via Debian)
Approved by: portmgr blanket
This allows to build lang/go14 on old x86 CPUs without SSE2 instructions.
This has an impact on generated machine code performance, but lang/go14
is only useful to boostrap lang/go so it is ok to do so unconditionally.
PR: 230733
Submitted by: Dmitri Goutnik <dg@syrec.org>
- Cleanup of variables
- Use the 'bsd' target for build stage
- Add options to to select interactive command line editing
- Add options to control debug options
- Add option for HTML documentation
- No change log provided upstream
Submitted by: Russel Haley (russ.haley@gmail.com)
Andrew Gierth (andrew_tao173.riddles.org.uk)
Differential Revision: https://reviews.freebsd.org/D13690
Vala 0.36.15
released on Jul 16, 2018
Changes
* Various improvements and bug fixes:
+ Update bugtracker links to
https://gitlab.gnome.org/GNOME/vala/issues
+ codegen:
- Handle delegate_target attribute of fields [#520]
- Free generic elements of glib collections [#694765]
- Fix warning for source_funcs parameter of g_source_new()
- Actually treat GLib.Source as compact class
- Consistently use gpointer for delegate targets
- Unify some delegate-type check patterns
+ girparser:
- Add support for bool delegate_target [#520]
- Allow changing the type of constants using metadata [#643]
+ girwriter:
- Output deprecated="1"
- construct-only properties don't have a setter method
- Don't write any custom attributes [#608]
- Add boolean GIR visible attribute to allow skipping symbols
+ Fix links in the Manual
+ Allow TypeModule subclasses as parameter-type in module-init-method
[#648]
+ vala: Report error for error-domains without any code
+ parser: Accept "empty" enums/error-domains [#253]
+ codewriter: Output bodies of code-blocks if in VAPIGEN mode [#253]
* Bindings:
+ glib-2.0:
- Do not mark simple-type out-parameters as nullable [#634]
- Make Source.attach() use the default MainContext by default
- Make Bytes.slice() use the memory-efficient Bytes.from_bytes()
[#638]
- Fix Bytes.with_free_func()
+ gio-2.0:
- Mark ActionEntry.parameter_type/state fields as nullable [#632]
- Delegate fields of DBus*VTable do not have implicit target fields
- Apply delegate_target = false on ActionEntry callbacks and make
them weak [#630]
- Add default to all io_priority parameters
- Add custom IOError.from_errno() to make it work as expected [#253]
+ gobject-2.0: Add lower_case_csuffix attributes to Type* symbols
+ gstreamer-rtsp-server-1.0: Update from 1.12+ git master [#653]
Vala 0.36.14
released on May 22, 2018
Changes
* Various improvements and bug fixes:
+ Warn if --gir is used without --library [#708693]
+ Report an error if rank attribute is required but was not found
[#660991]
+ scanner: Accept \R and \N escape sequences in regex literals
[#749576]
+ scanner: Accept backreferences \1 through \99 in regex literals
[#728574]
+ codegen:
- Avoid critical in get_basic_type_info() when there is no signature
[#729517]
- Fix serializing multidimensional arrays [#735437]
- Guard get_default_ref_sink_function() against null base-class
- Respect cname ccode-attribute of lockable fields
- Fix casting of length for "ref" array parameters
+ girparser: Add support for NoAccessorMethod metadata
+ girparser: Deal with missing "name" attribute by falling back on
"c:type"
+ girwriter: Drop errornous indent
+ girwriter: Properly output variadic methods and use
introspectable="0"
+ genie: Add support for the \v escape character [#664689]
+ genie: Add support for the \uXXXX escape sequence
* Bindings:
+ glib-2.0: Fix IConv.iconv() [#634753]
+ glib-2.0: Mark str parameter of Variant.take_string() as owned
+ gobject-2.0: Make constructor of ParamSpecBoxed public
+ gstreamer: Cherry-pick some fixes from 0.42
+ gtk+-*.0: Set NoAccessorMethod on Gtk.Radio*.group properties
+ json-glib-1.0: Mark return-type of gvariant_deserialize*() as
nullable [#795265]
+ sqlite3: Add additional constants for flags of Database.open_v2()
[#795627]
+ webkit2gtk-4.0: Update to 2.20.1 and cherry-pick some fixes from
0.42
Also,
Fix the problem of some non-header files as headers (Thanks to yuri@)
Fix Jupyter integration
PR: 230637
Reported by: yuri
Sponsored by: The FreeBSD Foundation
Details:
- Update to 0.24.4
- Avoid using -march=native
- Add pkg-message to let the users know that devel/pcre2 is neded
for regex/glob support in pony programs
PR: 230422
Submitted by: greg@unrelenting.technology (mantainer)
- Unfortunately, it now requires GNU make to build (pattern rules
in doc/Makefile.in); also, regression tests fail with BSD make:
https://github.com/shirok/Gauche/issues/367
Currently the gd-module uses a bundled libgd, while most systems
already provide the same library via graphics/gd.
Therefore instead of adding the bundled library we use the
port instead.
PR: 217222
Submitted by: Mikhail Teterin <mi@FreeBSD.org>
Currently the gd-module uses a bundled libgd, while most systems
already provide the same library via graphics/gd.
Therefore instead of adding the bundled library we use the
port instead.
PR: 217222
Submitted by: Mikhail Teterin <mi@FreeBSD.org>
Currently the gd-module uses a bundled libgd, while most systems
already provide the same library via graphics/gd.
Therefore instead of adding the bundled library we use the
port instead.
PR: 217222
Submitted by: Mikhail Teterin <mi@FreeBSD.org>
Those two ports fail to build on FreeBSD/i386 (all versions) for quite
some time. The Erlang/OTP team provided a patch after Dewayne Geraghty
reported the issue upstream.
Here is the entry in the Erlang/OTP bug tracker:
https://bugs.erlang.org/browse/ERL-677
The patch was successfully tested on FreeBSD 10.4 and 11.1, both i386
and amd64. Dewayne also confirmed success on FreeBSD 11.2.
PR: 229600
Reported by: Dewayne Geraghty <dewayne@heuristicsystems.com.au>
Obtained from: https://bugs.erlang.org/browse/ERL-677
MFH: 2018Q3
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
in preparation of the cycle leading up to the GCC 8.3 release in a few
months.
Functionally this is pretty much equivalent to the 8.2 release for the
time being.
lang/ruby23 is already fixed same way, port it to newer releases
By the way, this unbreak ninja builds with any port relied on libruby.so
PR: 229898
Submitted by: fluffy
Reviewed by: ruby (miwi)
Approved by: ruby (miwi)
Exp-run by: antoine
MFH: 2018Q3
Differential Revision: D16341
This fixes more than 99 bugs, among others a performance regression
with -mtune=native on Intel Skylake chips has been fixed.
This also fixes a bug introduced in GCC 8.1 affecting passing or
returning of classes with a deleted copy constructor and defaulted
trivial move constructor (GCC bug c++/86094). GCC 8.2 introduces
-fabi-version=13 and makes it the default; ABI incompatibilities
between GCC 8.1 and 8.2 can be reported with -Wabi=12.
