smlnj uses clang and inconsistent flags in Rosetta, fails to build

[barracuda156]

/usr/bin/clang -m64 -std=c99 -c -g -O2 -D_DARWIN_C_SOURCE -Os -arch ppc -DARCH_AMD64 -DSIZE_64 -DOPSYS_UNIX -DOPSYS_DARWIN -DGNU_ASSEMBLER -DDLOPEN -DINDIRECT_CFUNC -I../objs -I../include ../kernel/error.c
clang: warning: not using the clang compiler for the 'powerpc' architecture
/usr/bin/clang -m64 -std=c99 -c -g -O2 -D_DARWIN_C_SOURCE -Os -arch ppc -DARCH_AMD64 -DSIZE_64 -DOPSYS_UNIX -DOPSYS_DARWIN -DGNU_ASSEMBLER -DDLOPEN -DINDIRECT_CFUNC -I../objs -I../include ../kernel/timers.c
clang: warning: not using the clang compiler for the 'powerpc' architecture
/usr/bin/clang -m64 -std=c99 -c -g -O2 -D_DARWIN_C_SOURCE -Os -arch ppc -DARCH_AMD64 -DSIZE_64 -DOPSYS_UNIX -DOPSYS_DARWIN -DGNU_ASSEMBLER -DDLOPEN -DINDIRECT_CFUNC -I../objs -I../include ../kernel/unix-timers.c
clang: warning: not using the clang compiler for the 'powerpc' architecture
/usr/bin/clang -m64 -std=c99 -c -g -O2 -D_DARWIN_C_SOURCE -Os -arch ppc -DARCH_AMD64 -DSIZE_64 -DOPSYS_UNIX -DOPSYS_DARWIN -DGNU_ASSEMBLER -DDLOPEN -DINDIRECT_CFUNC -I../objs -I../include ../kernel/qualify-name.c
clang: warning: not using the clang compiler for the 'powerpc' architecture
/usr/bin/clang -m64 -std=c99 -c -g -O2 -D_DARWIN_C_SOURCE -Os -arch ppc -DARCH_AMD64 -DSIZE_64 -DOPSYS_UNIX -DOPSYS_DARWIN -DGNU_ASSEMBLER -DDLOPEN -DINDIRECT_CFUNC -I../objs -I../include ../kernel/swap-bytes.c
clang: warning: not using the clang compiler for the 'powerpc' architecture
/usr/bin/clang -m64 -std=c99 -c -g -O2 -D_DARWIN_C_SOURCE -Os -arch ppc -DARCH_AMD64 -DSIZE_64 -DOPSYS_UNIX -DOPSYS_DARWIN -DGNU_ASSEMBLER -DDLOPEN -DINDIRECT_CFUNC -I../objs -I../include ../mach-dep/unix-fault.c
clang: warning: not using the clang compiler for the 'powerpc' architecture
../mach-dep/unix-fault.c: In function ‘FaultHandler’:
../mach-dep/unix-fault.c:75: error: ‘struct __darwin_ppc_thread_state’ has no member named ‘__rip’
../mach-dep/unix-fault.c:88: warning: cast to pointer from integer of different size
../mach-dep/unix-fault.c:102: error: ‘struct __darwin_ppc_thread_state’ has no member named ‘__rip’
../mach-dep/unix-fault.c:102: warning: cast from pointer to integer of different size
make[1]: *** [unix-fault.o] Error 1
make: *** [all] Error 2
/opt/local/var/macports/build/_opt_local_var_macports_sources_rsync.macports.org_macports_release_tarballs_ports_lang_smlnj/smlnj/work/smlnj-110.99.3/config/install.sh: !!! Run-time system build failed for some reason.
Command failed:  cd "/opt/local/var/macports/build/_opt_local_var_macports_sources_rsync.macports.org_macports_release_tarballs_ports_lang_smlnj/smlnj/work/smlnj-110.99.3" && /opt/local/var/macports/build/_opt_local_var_macports_sources_rsync.macports.org_macports_release_tarballs_ports_lang_smlnj/smlnj/work/smlnj-110.99.3/config/install.sh 
Exit code: 1

[kencu (Ken)]

why is this port selecting "clang" to use here, when building for PPC?

It would be very nice if we could stop that from happening centrally, rather than continue to add hundreds of workarounds for this issue in individual Portfiles that will all need to be stripped out again if we get this fixed.

[kencu (Ken)]

this particular port seems to need some love anyway, as it is putting "-m64" and "-arch ppc" and "-DARCH_AMD64" on the same build line...

[tobypeterson]

Yeah, this port has major issues right now. I can't even build it properly on Apple silicon. Could be related to the wonky flags.