You raised that issue in this ticket
https://trac.macports.org/ticket/71601#comment:3 and I mentioned there I had also noted that error here: https://trac.macports.org/ticket/71503#comment:5 So I think we can make https://trac.macports.org/ticket/71601 about this. And it may turn out to be a problem, indeed, although when I raised it, nobody else said at that time they had seen it. Ken > On Dec 19, 2024, at 3:41 PM, Riccardo Mottola <riccardo.mott...@libero.it> > wrote: > > Hi! > > I am churning updates on all my three 10.5 systems! > > I just finished compiling gcc 14 on my PPC today.. even if it is a 2x 1GHz G4 > with 2GB of RAM it took quite a bit. The first attempt froze the terminal, so > I restarted using xterm and it worked - maybe just by luck. > > It also proves that GCC is very slow on PPC, compared to the MacBook which is > nominally maybe 50% faster, not more > > Ken Cunningham wrote: >> The default gcc update to gcc14 for older systems appears to work on all the >> systems where it has been tested. Kudos to the folks who did the heavy >> lifting in the PR for that, and for feedback from Josh and Chris to help it >> go as smoothly as possible. > > now I have some fallout.. I started opening a bug, but then realized that > some packages refuse to build allo for the same reason: > > dyld: lazy symbol binding failed: Symbol not found: __ZNSo9_M_insertIeEERSoT_ > Referenced from: /opt/local/bin/cmake-orig > Expected in: /opt/local/lib/libgcc/libstdc++.6.dylib > > dyld: Symbol not found: __ZNSo9_M_insertIeEERSoT_ > Referenced from: /opt/local/bin/cmake-orig > Expected in: /opt/local/lib/libgcc/libstdc++.6.dylib > > /bin/sh: line 1: 5829 Trace/BPT trap "/opt/local/bin/cmake" > -DCMAKE_C_COMPILER="/usr/bin/gcc-4.2" -DCMAKE_C_FLAGS="-Os -arch ppc" > -DCMAKE_EXE_LINKER_FLAGS="-L/opt/local/lib -Wl,-headerpad_max_install_names > -arch ppc" -DCMAKE_VERBOSE_MAKEFILE=ON -DCMAKE_MAKE_PROGRAM=make > -DBUILD_SHARED_LIBS=OFF -DCMAKE_POSITION_INDEPENDENT_CODE=ON > -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=./installed . > make: *** [brotli/libbrotlienc.a] Error 133 > > > be it using gcc-4.2 or gcc14, the issue is in > /opt/local/lib/libgcc/libstdc++.6.dylib which is a fresh build! > > Ideas? On x86 I don't seem to have this issue. > > Do we have a specific issue where to track this fallout? > I fear that now compilation of any C++ port is broken! > > > Riccardo
