[Bug middle-end/54848] New: -ftrapv doesn't work when assigning to an integer with smaller size

["mikulas at artax dot karlin.mff.cuni.cz" <gcc-bugzilla at gcc dot gnu dot org>]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54848

             Bug #: 54848
           Summary: -ftrapv doesn't work when assigning to an integer with
                    smaller size
    Classification: Unclassified
           Product: gcc
           Version: 4.7.2
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: middle-end
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: mikulas@artax.karlin.mff.cuni.cz
              Host: i686-pc-linux-gnu
            Target: i686-pc-linux-gnu
             Build: i686-pc-linux-gnu


When adding, subtracting or multiplying two integers and assigning the result
to a variable with smaller size, trap on overflow doesn't happen.

See for example this program:

#include <stdio.h>

int main(void)
{
        int a = 0x70000000;
        int b = 0x50000000;
        short c = a + b;
        printf("%x\n", c);
        return 0;
}

We compile the program without optimization and with -ftrapv, so that it should
crash because of integer overflow. But the program doesn't crash. If we replace
"short" with "int", an overflow crash happens correctly.