[Users] ETK config-stage failure with Macports gcc -- C+11 range-based for statments?

Erik Schnetter schnetter at gmail.com
Fri Nov 17 08:26:13 CST 2023 

In Cactus, "LD" needs to point to a compiler driver (e.g. "g++"), not
to a linker. You should see a very different error message when you
set "LD" to a linker. (Or maybe the configure script overwrites "LD".)

Can you try installing a different "ld64" package, installing a
different variant, or uninstalling it?

-erik

On Thu, Nov 16, 2023 at 5:29 PM Bernard Kelly <physicsbeany at gmail.com> wrote:
>
> Thanks, Erik.
>
> Hmm, OK. It turns out that I *have* ld64 installed via MacPorts, and I think that's what's being used here:
>
> gs66-shannon:~ bjkelly1$ L /opt/local/bin/ld*
> lrwxr-xr-x 1 root admin    8 Sep 28 19:42 /opt/local/bin/ld -> ld-xcode
> lrwxr-xr-x 1 root admin   16 Sep 28 19:42 /opt/local/bin/ld-classic -> ld-classic-xcode
> lrwxr-xr-x 1 root admin   32 Sep 28 19:41 /opt/local/bin/ld-classic-xcode -> ../libexec/ld64/ld-classic-xcode
> lrwxr-xr-x 1 root admin   24 Sep 28 19:41 /opt/local/bin/ld-xcode -> ../libexec/ld64/ld-xcode
> -rwxr-xr-x 1 root admin  176 Sep 28 20:06 /opt/local/bin/ld.lld-mp-14
> -rwxr-xr-x 1 root admin  180 Sep 28 20:06 /opt/local/bin/ld64.lld-mp-14
>
> I just tried setting LD = /opt/local/bin/ld-classic , but the failure is the same.
>
> I then tried using the pre-MacPorts system linker, LD = /usr/bin/ld, but the failure is unchanged.
>
> Perhaps there's a linker flag I should be setting? Right now I have nothing in LDFLAGS.
>
> Bernard
>
> On Thu, Nov 16, 2023 at 4:41 PM Erik Schnetter <schnetter at gmail.com> wrote:
>>
>> Bernard
>>
>> Your C++ compiler is probably fine, but your linker ("ld") is not, it
>> cannot handle GCC's output.
>>
>> I have a MacPorts package "ld64" installed. Maybe that would help?
>>
>> -erik
>>
>> On Thu, Nov 16, 2023 at 3:49 PM Bernard Kelly <physicsbeany at gmail.com> wrote:
>> >
>> > Hi Roland.
>> >
>> > I'm attaching the config.log file here (I had to hunt it down in configs/sim/config-data, since the configure step failed before it could point to anything).
>> >
>> > Since the logfile gives a small C++ program that failed to compile, I copied it to testGPP2.cpp, and tried compiling it myself; and yes, it failed. Errors from this are short enough to quote here:
>> >
>> > ------------
>> > gs66-shannon:Cactus_ET_2023_05 bjkelly1$ /opt/local/bin/g++-mp-13 testGPP2.cpp
>> > ld: warning: ignoring duplicate libraries: '-lgcc'
>> > 0  0x1049cb648  __assert_rtn + 72
>> > 1  0x1048fffac  ld::AtomPlacement::findAtom(unsigned char, unsigned long long, ld::AtomPlacement::AtomLoc const*&, long long&) const + 1204
>> > 2  0x104915924  ld::InputFiles::SliceParser::parseObjectFile(mach_o::Header const*) const + 15164
>> > 3  0x104922e30  ld::InputFiles::parseAllFiles(void (ld::AtomFile const*) block_pointer)::$_7::operator()(unsigned long, ld::FileInfo const&) const + 420
>> > 4  0x1a0274440  _dispatch_client_callout2 + 20
>> > 5  0x1a0289544  _dispatch_apply_invoke_and_wait + 224
>> > 6  0x1a028884c  _dispatch_apply_with_attr_f + 1180
>> > 7  0x1a0288a38  dispatch_apply + 96
>> > 8  0x10499d3b8  ld::AtomFileConsolidator::parseFiles(bool) + 292
>> > 9  0x10493e170  main + 9048
>> > ld: Assertion failed: (resultIndex < sectData.atoms.size()), function findAtom, file Relocations.cpp, line 1336.
>> > collect2: error: ld returned 1 exit status
>> > ------------
>> >
>> > As for your other question, no, there's no environment variable set with a name containing CXX.
>> >
>> > Thanks, Bernard
>> >
>> >
>> > On Thu, Nov 16, 2023 at 3:35 PM Roland Haas <rhaas at illinois.edu> wrote:
>> >>
>> >> Hello Bernard,
>> >>
>> >> Odd. You have the correct CXXFLAGS and the compiler seems new enough.
>> >>
>> >> Could you attach the config.log file that the run produced (should be
>> >> mentioned near the end with full path).
>> >>
>> >> Since this has been an issue in the past: there are no environment
>> >> variables CXX or CXXFLAGS defined in you shell (env | grep CXX finds
>> >> nothing)?
>> >>
>> >> Yours,
>> >> Roland
>> >>
>> >> > Hi. I tried making a new configuration on my Macbook (MacOS Ventura,
>> >> > 13.6.2) after a gap of several months, and it's now failing the "make
>> >> > sim-config" step with a complaint about the C++ compiler not being
>> >> > C++11-compliant -- end of the make config step screen output is below.
>> >> >
>> >> > I'm using the latest available Macports-supplied GNU compilers
>> >> > (gcc-mp-13, g++-mp-13, etc.), with the following C & C++ options:
>> >> >
>> >> > ---------- compiler flags -----------
>> >> > CPP = /opt/local/bin/cpp-mp-13
>> >> > CC  = /opt/local/bin/gcc-mp-13
>> >> > CXX = /opt/local/bin/g++-mp-13
>> >> > CPPFLAGS =
>> >> > CFLAGS   = -std=c99
>> >> > CXXFLAGS = -std=c++14
>> >> > C_LINE_DIRECTIVES = yes
>> >> > OPTIMISE           = yes
>> >> > CPP_OPTIMISE_FLAGS = # -DCARPET_OPTIMISE -DNDEBUG
>> >> > C_OPTIMISE_FLAGS   = -O2 -march=native
>> >> > CXX_OPTIMISE_FLAGS = -O2 -march=native
>> >> > OPENMP           = yes
>> >> > CPP_OPENMP_FLAGS = -fopenmp
>> >> > C_OPENMP_FLAGS   = -fopenmp
>> >> > CXX_OPENMP_FLAGS = -fopenmp
>> >> > WARN           = yes
>> >> > CPP_WARN_FLAGS = -Wall
>> >> > C_WARN_FLAGS   = -Wall
>> >> > CXX_WARN_FLAGS = -Wall
>> >> > ---------- end of compiler flags -----------
>> >> >
>> >> > Can anyone suggest what's up? The error says that the C++ compiler is
>> >> > failing to do range-based for statements, but I can compile & run a simple
>> >> > program with a range-based for statement, so this seems unlikely.
>> >> >
>> >> > I'm currently on ET_2023-05, but have this problem with ET_2022_11 as well,
>> >> > and with gcc-mp-12.
>> >> >
>> >> > Thanks, Bernard
>> >> >
>> >> > --------------- last bit of screen output ---------
>> >> > checking if compiler has broken omp collapse... no
>> >> > checking for getopt_long_only... yes
>> >> > checking for working const... yes
>> >> > checking for C inline... inline
>> >> > checking for C static inline... static inline
>> >> > checking for C restrict... restrict
>> >> > checking for C++ restrict... __restrict__
>> >> > checking for C++ copysign... std::copysign
>> >> > checking for C++ fpclassify... std::fpclassify
>> >> > checking for C++ isfinite... std::isfinite
>> >> > checking for C++ isinf... std::isinf
>> >> > checking for C++ isnan... std::isnan
>> >> > checking for C++ isnormal... std::isnormal
>> >> > checking for C++ signbit... std::signbit
>> >> > checking for C _Pragma... yes
>> >> > checking for C function __attribute__((__const__))... yes
>> >> > checking for C++ function __attribute__((__const__))... yes
>> >> > checking for C++ member function __attribute__((__const__))... yes
>> >> > checking for C function __attribute__((__pure__))... yes
>> >> > checking for C++ function __attribute__((__pure__))... yes
>> >> > checking for C++ member function __attribute__((__pure__))... yes
>> >> > checking for C data __attribute__((__common__))... yes
>> >> > checking for C+ data __attribute__((__common__))... yes
>> >> > checking for C function __attribute__((__noinline__))... yes
>> >> > checking for C++ function __attribute__((__noinline__))... yes
>> >> > checking for C++ member function __attribute__((__noinline__))... yes
>> >> > checking for C function __attribute__((__always_inline__))... yes
>> >> > checking for C++ function __attribute__((__always_inline__))... yes
>> >> > checking for C++ member function __attribute__((__always_inline__))... yes
>> >> > checking for C __attribute__((__unused__))... yes
>> >> > checking for C++ __attribute__((__unused__))... yes
>> >> > checking for C __attribute__((__aligned__(...)))... yes
>> >> > checking for C++ __attribute__((__aligned__(...)))... yes
>> >> > checking for C __attribute__((__cold__))... yes
>> >> > checking for C++ __attribute__((__cold__))... yes
>> >> > checking for C __attribute__((__hot__))... yes
>> >> > checking for C++ __attribute__((__hot__))... yes
>> >> > checking for C __attribute__((__format__(printf, 1, 2)))... yes
>> >> > checking for C++ __attribute__((__format__(printf, 1, 2)))... yes
>> >> > checking for C __attribute__((__noreturn__))... yes
>> >> > checking for C++ __attribute__((__noreturn__))... yes
>> >> > checking for C __attribute__((__nonnull__))... yes
>> >> > checking for C++ __attribute__((__nonnull__))... yes
>> >> > checking for C __attribute__((__returns_nonnull__))... yes
>> >> > checking for C++ __attribute__((__returns_nonnull__))... yes
>> >> > checking for C __builtin_expect... yes
>> >> > checking for C++ __builtin_expect... yes
>> >> > checking for C __builtin_trap... yes
>> >> > checking for C++ __builtin_trap... yes
>> >> > checking for C __builtin_unreachable... no
>> >> > checking for C++ __builtin_unreachable... no
>> >> > checking for C __builtin_assume_aligned... yes
>> >> > checking for C++ __builtin_assume_aligned... yes
>> >> > checking for C++ static_assert... yes
>> >> > checking for C++ auto specifier... yes
>> >> > checking for C++ lambda expressions... yes
>> >> > checking for C++ range-based for statements... no
>> >> > Cactus requires a C++11 compiler -- check your C++ compiler and C++
>> >> > compiler flags
>> >> >
>> >> > Error reconfiguring sim-config
>> >> > make: *** [sim-config] Error 2
>> >> >
>> >>
>> >>
>> >> Yours,
>> >> Roland
>> >>
>> >> --
>> >> My email is as private as my paper mail. I therefore support encrypting
>> >> and signing email messages. Get my PGP key from http://pgp.mit.edu .
>> >
>> >
>> >
>> > --
>> > ------------------------------------------------------------------
>> > Bernard Kelly -- CRESST Assistant Research Scientist, NASA/GSFC
>> > Gravitational Astrophysics Laboratory --- Code 663
>> >
>> > Phone: +1 (301) 286-7243 *** Fax: +1 (301) 286-2226
>> > Web: http://science.gsfc.nasa.gov/sed/bio/bernard.j.kelly
>> > ORCID: orcid.org/0000-0002-3326-4454
>> > ------------------------------------------------------------------
>> > _______________________________________________
>> > Users mailing list
>> > Users at einsteintoolkit.org
>> > http://lists.einsteintoolkit.org/mailman/listinfo/users
>>
>>
>>
>> --
>> Erik Schnetter <schnetter at gmail.com>
>> http://www.perimeterinstitute.ca/personal/eschnetter/
>
>
>
> --
> ------------------------------------------------------------------
> Bernard Kelly -- CRESST Assistant Research Scientist, NASA/GSFC
> Gravitational Astrophysics Laboratory --- Code 663
>
> Phone: +1 (301) 286-7243 *** Fax: +1 (301) 286-2226
> Web: http://science.gsfc.nasa.gov/sed/bio/bernard.j.kelly
> ORCID: orcid.org/0000-0002-3326-4454
> ------------------------------------------------------------------



-- 
Erik Schnetter <schnetter at gmail.com>
http://www.perimeterinstitute.ca/personal/eschnetter/

More information about the Users mailing list