Mysterious Branch Failures with PPC/Darwin Make

[James E Keenan]

More data.

Good news: It's not a problem strictly associated with plobsing's  
branch.  Bad news: It's found in master as well.

Specifically, this problem seems to crop up when I do an "all g++"  
build on Darwin/PPC (as distinct from my customary cc=gcc and g++ for  
link and ld).

$ git branch | grep \*;git show | head -1;grep -E "'(cc|link|ld)'"  
lib/Parrot/Config/Generated.pm
* master
commit 49f46ec975761bc5010ef4d65dc30fb5651bd77e
   'cc' => 'g++',
   'ld' => '/usr/bin/g++',
   'link' => '/usr/bin/g++',

src/gc/gc_gms.c:1231: failed assertion '!PObj_on_free_list_TEST(pmc)  
|| !"Resurrecting of dead objects is not supported"'
make: *** [compilers/opsc/gen/Ops/Trans/C.pir] Abort trap
./parrot-nqp --target=pir --output=compilers/opsc/gen/Ops/Trans/C.pir  
compilers/opsc/src/Ops/Trans/C.pm
src/gc/gc_gms.c:1231: failed assertion '!PObj_on_free_list_TEST(pmc)  
|| !"Resurrecting of dead objects is not supported"'
make: *** [compilers/opsc/gen/Ops/Trans/C.pir] Abort trap

But one new aspect:  When I tried to run the failed command by  
itself, 'make' did not continue because the command itself failed.

$ ./parrot-nqp --target=pir --output=compilers/opsc/gen/Ops/Trans/ 
C.pir compilers/opsc/src/Ops/Trans/C.pm
PARROT VM: Could not initialize new interpreterunimplemented value type

kid51