ba445bc0-1011-4361-8560-c5d362a94dfe
commented
4 years ago This also occurs with std::variant in Clang++. Something about libstdc++'s std::variant jogs GCC's memory, though; it works in GCC.
// Tested in whatever "x86-64 clang (trunk)" is on Godbolt.
//clang++ -std=gnu++17 -O3 -fstrict-aliasing meow.cpp -o meow
#include <variant>
#include <cstdio>
struct s1 {unsigned short x;};
struct s2 {unsigned short x;};
using s1s2 = std::variant<s1, s2>;
static int read_s1x(s1 *p) { return p->x; }
static void write_s2x(s2 *p, int v) { p->x=v; }
int test(s1s2 *p1, s1s2 *p2, s1s2 *p3)
{
if (read_s1x(std::get_if<s1>(p1)))
{
unsigned short temp;
temp = std::get_if<s1>(p3)->x;
p3->emplace<s2>(s2{temp});
write_s2x(std::get_if<s2>(p2),1234);
temp = std::get_if<s2>(p3)->x;
p3->emplace<s1>(s1{temp});
}
return read_s1x(std::get_if<s1>(p1));
}
int test2(int x)
{
s1s2 q[2];
q->emplace<s1>(s1{4321});
return test(q,q+x,q+x);
}
int main(void)
{
std::printf("%d\n",test2(0));
}
hfinkel
davidstone
llvmbot
ec04fc15-fa35-46f2-80e1-5d271f2ef708
Extended Description
Consider the following C/C++ code with -O3 -fstrict-aliasing:
Clang (and GCC) generate code that outputs 4321 instead of the expected 1234.
I don't really understand things in terms of the C standard, but in terms of the C++ standard, it seems as if Clang and GCC are incorrect, and this code is well-defined. (The output is 4321 in both C and C++ mode.)
According to [class.union]/5 in the C++17 draft N4659, the assignment expression "p3->v2.x = temp;" changes the active member of the union. It's done through a union member access expression. Thus the pointer &p2->v2 is valid here.
Even if I switch this to "p3->v2 = { x };", avoiding the nested case, the problem still happens.
Even if I explicitly change the active member of the union with placement new as "new(&p3.v2) s2;", the problem still happens.
Is it possible that Clang doesn't see the possibility that p2 and p3 point to the same object?