On Dec 11, 2006, at 9:42 AM, jerry gay wrote:

>  On 12/10/06, James E Keenan <jkeen_via_google@[...].com> wrote:
> > James E Keenan wrote:
> >
> > And now here's what I got with
> >
> > perl Configure.pl  (i.e., no options)
> > vi Makefile (to eliminate -bundle from LD_LOAD_FLAGS)
> > make
> >
> > Failure at this point has now become frequently observed.
> >
> >
> > [parrot] 631 $ perl Configure.pl
> > [parrot] 632 $ vi Makefile # to delete -bundle from LD_LDOAD_FLAGS
> > [parrot] 633 $ make
> 
>  <snip most of configure/make output>
> 
> > c++ -dynamiclib -undefined suppress -L/usr/local/lib -L/Users/jimk/ 
> > work/parrot/blib/lib -flat_namespace   -o blib/lib/libparrot. 
> > 0.4.7.dylib -install_name /usr/local/lib/libparrot.dylib >>     src/ops/core_ops.o src/op
s/core_ops_switch.o src/asmfun.o src/ 
> > ops/core_ops_cg.o src/ops/core_ops_cgp.o src/exec.o src/exec_cpu.o  
> > src/exec_save.o src/charset.o src/encoding.o src/exceptions.o src/ 
> > global_setup.o src/interpreter.o src/inter_call.o src/inter_cb.o  
> > src/inter_create.o src/inter_misc.o src/inter_run.o src/register.o  
> > src/memory.o src/objects.o src/packfile.o src/stacks.o src/ 
> > stack_common.o src/string.o src/string_primitives.o src/sub.o src/ 
> > runops_cores.o src/trace.o src/pmc.o src/pmc_freeze.o src/key.o  
> > src/hash.o src/core_pmcs.o src/platform.o src/jit.o src/jit_cpu.o  
> > src/jit_debug.o src/jit_debug_xcoff.o src/intlist.o src/list.o src/ 
> > library.o src/global.o src/embed.o src/warnings.o src/packout.o  
> > src/byteorder.o src/debug.o src/smallobject.o src/headers.o src/ 
> > hll.o src/dod.o src/gc_ims.o src/gc_gms.o src/exit.o src/misc.o  
> > src/spf_render.o src/spf_vtable.o src/datatypes.o src/ 
> > fingerprint.o src/nci.o src/cpu_dep.o src/tsq.o src/longopt.o src/ 
> > events.o src/thread.o src/dynext.o src/utils.o src/vtables.o src/ 
> > pic.o src/pic_jit.o src/mmd.o src/builtin.o src/extend.o src/ 
> > extend_vtable.o src/revision.o src/packfile/pf_items.o src/stm/ 
> > backend.o src/stm/waitlist.o src/resources.o src/charset/ascii.o  
> > src/charset/binary.o src/charset/iso-8859-1.o src/charset/tables.o  
> > src/charset/unicode.o src/io/io.o src/io/io_buf.o src/io/ 
> > io_layers.o src/io/io_unix.o src/io/io_utf8.o src/io/io_mmap.o src/ 
> > io/io_win32.o src/io/io_stdio.o src/io/io_string.o src/io/ 
> > io_passdown.o src/pmc/default.o src/pmc/null.o src/pmc/env.o src/ 
> > pmc/key.o src/pmc/random.o src/pmc/unmanagedstruct.o src/pmc/ 
> > managedstruct.o src/pmc/delegate.o src/pmc/csub.o src/pmc/ 
> > compiler.o src/pmc/exception.o src/pmc/version.o src/pmc/ 
> > vtablecache.o src/pmc/parrotio.o src/pmc/parrotlibrary.o src/pmc/ 
> > parrotinterpreter.o src/pmc/parrotthread.o src/pmc/lexpad.o src/ 
> > pmc/timer.o src/pmc/pointer.o src/pmc/sub.o src/pmc/closure.o src/ 
> > pmc/continuation.o src/pmc/retcontinuation.o src/pmc/ 
> > exception_handler.o src/pmc/coroutine.o src/pmc/eval.o src/pmc/ 
> > nci.o src/pmc/float.o src/pmc/integer.o src/pmc/bigint.o src/pmc/ 
> > complex.o src/pmc/string.o src/pmc/boolean.o src/pmc/ref.o src/pmc/ 
> > sharedref.o src/pmc/array.o src/pmc/fixedintegerarray.o src/pmc/ 
> > intlist.o src/pmc/iterator.o src/pmc/sarray.o src/pmc/ 
> > fixedstringarray.o src/pmc/multiarray.o src/pmc/hash.o src/pmc/ 
> > orderedhash.o src/pmc/tqueue.o src/pmc/parrotclass.o src/pmc/ 
> > parrotobject.o src/pmc/os.o src/pmc/file.o src/pmc/addrregistry.o  
> > src/pmc/bound_nci.o src/pmc/capture.o src/pmc/deleg_pmc.o src/pmc/ 
> > enumerate.o src/pmc/fixedbooleanarray.o src/pmc/fixedfloatarray.o  
> > src/pmc/fixedpmcarray.o src/pmc/lexinfo.o src/pmc/multisub.o src/ 
> > pmc/namespace.o src/pmc/pair.o src/pmc/parrotrunningthread.o src/ 
> > pmc/resizablebooleanarray.o src/pmc/resizablefloatarray.o src/pmc/ 
> > resizableintegerarray.o src/pmc/resizablepmcarray.o src/pmc/ 
> > resizablestringarray.o src/pmc/scalar.o src/pmc/slice.o src/pmc/ 
> > stmlog.o src/pmc/stmref.o src/pmc/stmvar.o src/pmc/super.o src/pmc/ 
> > undef.o src/encodings/fixed_8.o src/encodings/ucs2.o src/encodings/ 
> > utf16.o src/encodings/utf8.o compilers/imcc/imcparser.o compilers/ 
> > imcc/imclexer.o compilers/imcc/imc.o compilers/imcc/symreg.o  
> > compilers/imcc/instructions.o compilers/imcc/cfg.o compilers/imcc/ 
> > reg_alloc.o compilers/imcc/sets.o compilers/imcc/debug.o compilers/ 
> > imcc/optimizer.o compilers/imcc/pbc.o compilers/imcc/parser_util.o  
> > compilers/imcc/pcc.o -lm -lgmp
> > ld: common symbols not allowed with MH_DYLIB output format with  
> > the -multi_module option
> > /usr/local/lib/libgmp.a(mp_bpl.o) definition of common ___gmp_junk  
> > (size 4)
> > /usr/bin/libtool: internal link edit command failed
> > make: *** [blib/lib/libparrot.0.4.7.dylib] Error 1
> > [parrot] 634 $
> >
>  i note two things here.
>  1) you're using a c++ compiler to perform linking here. that's
>  destined for failure, and is why C<--cc=gcc> has been suggested as a
>  C<Configure.pl> option.

No, this is correct on OS X. Here's my shell script for running config:

#!/bin/sh
CC="ccache gcc-4.0"
CX="ccache g++-4.0"
/usr/bin/perl Configure.pl --cc="$CC" --cxx="$CX" --link="$CX" -- 
ld="$CX" --without-icu $@

You can s/ccache //g for the same effect sans without caching.

Regards.

>  2) you're attempting to link with an installed, yet broken C<gmp>
>  implementation, which is why C<--without-gmp> was created in the first
>  place. unless you remove C<gmp> from your system, you'll need to
>  continue using the C<--without-gmp> option, since linking will fail
>  otherwise.
> 
>  therefore, i suggest you continue to use C<Configure.pl --without-gmp
>  --cc=gcc> every time you configure parrot.
> 
>  i can't speak to C<--ccflags=-DAN>, because i'm not familiar with the
>  options. however, i see in a previous message, you attempted to run
>  C<Configure.pl --without-gmp --cc=gcc> (as i suggest) and it gave you
>  errors during execution. those errors, during compiler detection,
>  point you to look at a temporary file called F<test.cco>. if you could
>  provide the contents of this file in a reply, it would greatly aid in
>  debugging, and may explain why C<--ccflags=-DAN> was suggested. (i'll
>  research that in the meantime.)
> 
>  ~jerry

After talking to James on #parrot, I think the root of his problem is  
that he's not using apple-standard gcc. I'd like to see the results  
of using an apple-installed version of gcc/g++.

--
Will "Coke" Coleda
will@coleda.com