Hi Jonathan, > On 21/06/18 16:49 +0200, Rainer Orth wrote: >>Hi Jonathan, >> >>> No objection to this patch, but I'll just note that we have >>> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81091 suggesting we >>> should use LFS for libstdc++ unconditionally. >> >>seems like a wise move to me. The libstdc++.so ABI didn't change on >>Solaris either (that possibility had caused concern for me initially); >>didn't check libstdc++fs.a though. > > Well the main reason that's only a static library for now is to allow > us to make ABI incompatible changes before we declare it stable and > add those symbols to libstdc++.so forever.
I suspected that much. However, once you're reasonable confident the interface is stable, usability would be much improved by moving into libstdc++.so: e.g. before Solaris 11.4 sendfile lives in a separate libsendfile. This can easily be dealt with as a shared library dependency, but is harder (or puts the burden on the user) for static libstdc++fs.a. One might thing about introducing libstdc++.spec which could better deal with issues like this, just like libgfortran.spec. I'd started to work on this long ago do get rid of the hardcoded -lrt in g++, but never completed it. Rainer -- ----------------------------------------------------------------------------- Rainer Orth, Center for Biotechnology, Bielefeld University