Help trying to get gcc47 to run on ppc32/Tiger.
darylew at mac.com
Thu Jun 28 20:11:45 PDT 2012
On Jun 24, 2012, at 2:41 PM, Jeremy Huddleston Sequoia wrote:
> On Jun 23, 2012, at 3:03 PM, Daryle Walker <darylew at mac.com> wrote:
>> I just wanted to update my ports, and I ended up losing both of my
>> C++0x/11 compilers. I have the least-supported system, a 32-bit
>> PowerPC Mac running Mac OS x 10.4.11.
>> I lost gcc46, but I fixed it by following a comment. I haven't
>> been so lucky with gcc47, and looking at ticket 34385, hope has
>> been all but given up. On my last attempt at a fix, it seems that
>> either I messed up copying the patch from the gcc45 attempts, or
>> the affected file is different between gcc45 and gcc47. It
>> doesn't help that I don't know what's wrong in the first place,
>> nor whose end (MacPorts or GNU) does the problem lay.
>> Any help on what's wrong with the patch? (Then I can try compiling.)
> If I recall, the patch is actually applied already to gcc's
> libffi. You need to check yourself to figure out what's going wrong.
I'm trying to figure stuff out; that's why I'm asking for help here.
I'm just a port user, and now I'm suddenly have to be a port maintainer.
There is a MacPort for libffi, but the GCC port doesn't reference it,
right? GCC uses a copy of libffi in its own sources, right? When a
port has to be built from source, a copy of that source will be on my
disk, right? How/where can I check if it's still around?
> You'll probably get more support on this configuration from GCC
> developers directly. I suggest you file a bug report at http://
Right now, I don't know what to tell them. How different is the
MacPorts' version of the source from the official version; they might
not want to debug something too modified?
>> (At least we should require (Apple's?) CLang to compile GCC-4.6
>> when it's available, and Apple's GCC 4.2 otherwise. Never use the
>> LLVM-GCC 4.2, and install Apple GCC 4.2 when the computer doesn't
>> come with it. I also wrote in GCC 4.7 to use Apple's GCC 4.2 for
> That should have nothing to do with it.
>> (In the worst case, since we did have earlier versions of GCC 4.7
>> working on my computer, is there any way I can get that old
>> version back?)
> I'd be interested to know how you got that old version to compile,
> given this bug. What changed to cause it to fail? My suspicion is
> that you have a newer cctools installed from MacPorts which is
> noticing this bug whereas older cctools let it slip through to
> become a runtime issue.
I don't know what change happened. Here's some quick run results:
[daryle]$ port installed gcc47
The following ports are currently installed:
gcc47 @4.7.0_2 (active)
[daryle]$ g++-mp-4.7 --version
g++-mp-4.7 (GCC) 4.7.0
Copyright (C) 2012 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR
[daryle]$ g++-mp-4.7 hello.cpp
dyld: Library not loaded: /opt/local/lib/libppl.9.dylib
Referenced from: /opt/local/libexec/gcc/ppc-apple-darwin8/4.7.0/
Reason: image not found
g++-mp-4.7: internal compiler error: Trace/BPT trap (program cc1plus)
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.
My current version is based on previous efforts with the maintainers
to make gcc47 work with PowerPC systems. Any updates have failed,
but something else successfully changed to make LibPPL missing. I
think that library is part of LibFFI; but I couldn't update LibPPL.
(MacPorts kept wanting to fix gcc47 first.) Could I somehow force
LibPPL to be installed, without doing a gcc47 check? Would that
help, since gcc47 seems to use a private LibFFI anyway?
(BTW, the maintainers "fixed" the problem by killing PowerPC builds.
I have to hack the Portfile to do any testing.)
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the macports-dev