search for: csetjmp

...unction call. Throughout my experimenting, I was told that setjmp could create fake entry points into a region of code and that might prevent code motion.What I found is something surprising, and probably is a misuse of setjmp but I couldn't find an explanation for it.Consider this:#include <csetjmp> std::jmp_buf jb;int main() {  int s = 1;  setjmp(jb);   if (s) {    s = 0;    std::longjmp(jb, 1);    return 2;  }  return 1;} One would expect that the load of s in the if condition is not optimized away (by being replaced with if(1)), but clang at -O3 (on linux) generates this:define signext...

...s/out-arm/include/c++/v1/./__tuple > -- Installing: > /home/cschen/test/libcxx-scripts/out-arm/include/c++/v1/./cstring > -- Installing: > /home/cschen/test/libcxx-scripts/out-arm/include/c++/v1/./ctgmath > -- Installing: > /home/cschen/test/libcxx-scripts/out-arm/include/c++/v1/./csetjmp > -- Installing: > /home/cschen/test/libcxx-scripts/out-arm/include/c++/v1/./initializer_list > -- Installing: > /home/cschen/test/libcxx-scripts/out-arm/include/c++/v1/./__functional_base_03 > -- Installing: > /home/cschen/test/libcxx-scripts/out-arm/include/c++/v1/./new > --...

On 10/3/14 10:16 AM, Dan Albert wrote: >> >> I try to build libc++ and libc++abi for host x86_64(linux) and target >> arm(linux) but fail. >> > > Failing in what way? If this isn't working out of the box, we've done > something wrong. Yeah, it would help to know more specifics about where you're getting stuck. > > jroelofs might know more... For