libgcc +universal failing to compile on 10.9.2 with a gfortran related internal compiler error

[bernstei]

I'm trying to get MacPorts 2.2.1 to install gcc48 +universal on a 10.9.2 machine, XCode 5.1, command line tools installed manually from commandline_tools_os_x_mavericks_for_xcodemarch_2014.dmg

I start with a completely clean installation of macports (i.e. rm -rf /opt/local, installing macports from the binary installer dmg), do a selfupdate, upgrade outdated, then "port install gcc48 +universal"

I get an error while building libgcc referring me to /opt/local/var/macports/logs/_opt_local_var_macports_sources_rsync.macports.org_release_tarballs_ports_lang_gcc48/libgcc/main.log

In there I see the apparent problem:

:info:build configure: error: GNU Fortran is not working; please report a bug in http://gcc.gnu.org/bugzilla, attaching /opt/local/var/macports/build/_opt_local_var_macports_sources_rsync.macports.org_release_tarballs_ports_lang_gcc48/libgcc/work/build/x86_64-apple-darwin13/libgfortran/config.log

In /opt/local/var/macports/build/_opt_local_var_macports_sources_rsync.macports.org_release_tarballs_ports_lang_gcc48/libgcc/work/build/x86_64-apple-darwin13/libgfortran/config.log I see the gfortran related error:

configure:15630: checking whether the GNU Fortran compiler is working
configure:15643: /opt/local/var/macports/build/_opt_local_var_macports_sources_rsync.macports.org_release_tarballs_ports_lang_gcc48/libgcc/work/build/./gcc/gfortran -B/opt/local/var/macports/build/_opt_local_var_macports
_sources_rsync.macports.org_release_tarballs_ports_lang_gcc48/libgcc/work/build/./gcc/ -B/opt/local/x86_64-apple-darwin13/bin/ -B/opt/local/x86_64-apple-darwin13/lib/ -isystem /opt/local/x86_64-apple-darwin13/include -is
ystem /opt/local/x86_64-apple-darwin13/sys-include    -c -g -O2  conftest.f >&5
f951: internal compiler error: Illegal instruction: 4

f951: internal compiler error: Abort trap: 6
gfortran: internal compiler error: Abort trap: 6 (program f951)
/opt/local/var/macports/build/_opt_local_var_macports_sources_rsync.macports.org_release_tarballs_ports_lang_gcc48/libgcc/work/gcc-4.8.2/libgfortran/configure: line 1893: 63346 Abort trap: 6           $FC -c $FCFLAGS $ac_fcflags_srcext conftest.$ac_ext 1>&5
configure:15643: $? = 134
configure: failed program was:
| 
|       program foo
|       real, parameter :: bar = sin (12.34 / 2.5)
|       end program foo
configure:15647: result: no
configure:15649: error: GNU Fortran is not working; please report a bug in http://gcc.gnu.org/bugzilla, attaching /opt/local/var/macports/build/_opt_local_var_macports_sources_rsync.macports.org_release_tarballs_ports_lang_gcc48/libgcc/work/build/x86_64-apple-darwin13/libgfortran/config.log

I'll attach these two log files for reference

[bernstei]

/opt/local/var/macports/build/_opt_local_var_macports_sources_rsync.macports.org_release_tarballs_ports_lang_gcc48/libgcc/work/build/x86_64-apple-darwin13/libgfortran/config.log

[mf2k (Frank Schima)]

Please do not Cc openmaintainer@… because it is not a valid email address.

[bernstei]

Sorry - it was there in the port info Maintainers line.

[mf2k (Frank Schima)]

That is simply a placeholder stating that the port is openmaintainer. Excluding it is stated in the ​New Ticket Guidelines, albeit towards the end of the section.

[bernstei]

/opt/local/var/macports/logs/_opt_local_var_macports_sources_rsync.macports.org_release_tarballs_ports_lang_gcc48/libgcc/main.log

[bernstei]

Apparently I didn't read that section carefully enough. I also just realized that main.log is huge, so I tried to replace it with a bzip2 version, but I can't figure out how to remove the original.

[cooljeanius (Eric Gallager)]

This is the actual error from the log:

:info:build checking how to hardcode library paths into programs... immediate
:info:build checking whether the GNU Fortran compiler is working... no
:info:build configure: error: GNU Fortran is not working; please report a bug in http://gcc.gnu.org/bugzilla, attaching /opt/local/var/macports/build/_opt_local_var_macports_sources_rsync.macports.org_release_tarballs_ports_lang_gcc48/libgcc/work/build/x86_64-apple-darwin13/libgfortran/config.log
:info:build make[1]: *** [configure-target-libgfortran] Error 1
:info:build make[1]: Leaving directory `/opt/local/var/macports/build/_opt_local_var_macports_sources_rsync.macports.org_release_tarballs_ports_lang_gcc48/libgcc/work/build'
:info:build make: *** [bootstrap] Error 2
:info:build make: Leaving directory `/opt/local/var/macports/build/_opt_local_var_macports_sources_rsync.macports.org_release_tarballs_ports_lang_gcc48/libgcc/work/build'
:info:build Command failed:  cd "/opt/local/var/macports/build/_opt_local_var_macports_sources_rsync.macports.org_release_tarballs_ports_lang_gcc48/libgcc/work/build" && /usr/bin/make -j7 -w bootstrap 
:info:build Exit code: 2

In which case that means that this is a duplicate of #40548 (just like #41753 and #42752 were), which means that is actually an issue with mpfr (#42732) and/or libmpc (#42733) that would have been caught if their testsuites had been run automatically (#42731)

[cooljeanius (Eric Gallager)]

Cc Me!

[ryandesign (Ryan Carsten Schmidt)]

Replying to noam.bernstein@…:

I also just realized that main.log is huge, so I tried to replace it with a bzip2 version, but I can't figure out how to remove the original.

Thanks, I've removed the original. (Only administrators can remove attachments.)

[ryandesign (Ryan Carsten Schmidt)]

Duplicate of #40548.