vgvassilev wrote: > > > Are we sure this is going to do what we think? > > > https://github.com/llvm/llvm-project/blob/a41bb71f2216cef08ab04f1d730ae1701c145f3c/libcxx/include/cstdlib#L127 > > > > > > https://github.com/gcc-mirror/gcc/blob/494d3c3faaee0dbde696ea334f8e242ae85ae2b5/libstdc%2B%2B-v3/include/c_compatibility/stdlib.h#L65 > > > > > > I am not sure if I understand the code you pointed to. That's one of the > > reasons I did not want to expand that patch beyond clang and llvm. > > Both libc++ and libstdc++ (which target POSIX systems) seem to have no code > that provides those thread-safety guarantees; they're exposing `::getevn` as > `std::getenv`. So I don't think this actually resolves any issues. > > CC @ldionne @philnik777 for more opinions
The issue I linked in the PR description empirically seems to prove that `std::getenv` works better in a thread-safe environment when they say that they had to move `::getenv`->`std::getenv` to get rid of crashes... https://github.com/llvm/llvm-project/pull/108529 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
