https://gcc.gnu.org/bugzilla/show_bug.cgi?id=90330
--- Comment #10 from Matt Thompson <matthew.thompson at nasa dot gov> --- (In reply to Iain Sandoe from comment #9) > (In reply to Matt Thompson from comment #8) > > > > I just tried to rebuild gcc 9.1.0 on my laptop, and I got the same error at > > the same point in make install. > > One possible (although I suspect unlikely) issue is a race condition when > instslling symlinks which is known to occur on AFPS (the system I use for > testing is HHFS+). > > does it reliably still fail if you try to restart the install? It does indeed. Very repeatable. > > > I'll try a few other experiments (remove /usr/local/bin from my path in case > > brew is causing an issue, though it didn't before). > > OK - when I build toolchains the environment has no macports/fink/homebrew > content so there's no possibility of interaction. I don't usually remove it from my path because it had never caused any issues before. Note that I specifically do not let brew install gcc (or MPI). I like that control. I'm currently running a "no /usr/local/bin" test so we shall see if it helps. (I will say, as an aside, that I also can never get *spack* to build GCC on my laptop. Any version. Never figured that out either which is weird because it is a build from source!) > > > > > Or should I be using one of my other compilers to build gcc? > > I have built with both clang (10.2 command line tools) and a toolchain based > on GCC 8.3 (although not a formal release) - both builds succeeded - I have > no access to my Darwin18 system right now, so not able to attempt to repeat > your exercise. > > Presumably, the 8.2 toolchains is correctly installed and has not been > moved, so that the shared libraries from that are on the correct paths (the > kind of error you are seeing could be the wrong dylib being found). I install every release to its own tree and control access through lmod, so I can test this pretty easily. When I start up, I just have basics on my system (though I do symlink some brew utils to my ~/bin like gsed->sed because I cannot handle BSD sed. Grah!) > > If you can find the exact command line that fails and repeat it, prepended > by DYLD_PRINT_LIBRARIES=1, that will show you which dylibs are loaded as the > command tries to execute (there might be quite a lot of output, so not > necessarily easy to interpret) I just tried this and nothing happens. I tried both "DYLD_PRINT_LIBRARIES=1 g++ ..." and "env DYLD_PRINT_LIBRARIES=1 g++ ..." No change in output. > > I have built > > gcc 8.2.0. Instead of using clang, I could try that...though I'm not sure > > how. Just pass in CC and CXX to configure? > > yes, and/or make sure that the GCC toolchain bin dir is before the system > ones in your path. I'll try this as well as soon as I can (I can maybe manage make -j4 on this laptop, so builds aren't quick!). Do I need to pass FC as well? Or does the gcc build with fortran not care?
