[LLVMbugs] [Bug 15199] New: Different type conversion in ternary operator between clang and gcc

[bugzilla-daemon at llvm.org]

http://llvm.org/bugs/show_bug.cgi?id=15199

            Bug ID: 15199
           Summary: Different type conversion in ternary operator between
                    clang and gcc
           Product: clang
           Version: unspecified
          Hardware: PC
                OS: Linux
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: C++
          Assignee: unassignedclangbugs at nondot.org
          Reporter: klimek at google.com
                CC: dgregor at apple.com, llvmbugs at cs.uiuc.edu
    Classification: Unclassified

Consider the following snippet:
#include <cstdint>
#include <stdio.h>

const  int64_t kint64max  = ((int64_t)0x7FFFFFFFFFFFFFFFLL);

int main() {
  double d = (double)kint64max;
  int64_t v = d < kint64max ? d : kint64max;
  printf("%.20f -> %ld\n", d, v);
  return 0;
}

/tmp$ ~/build/llvm-opt/bin/clang++ -std=c++11 -o tt t.cc  && ./tt
9223372036854775808.00000000000000000000 -> -9223372036854775808
/tmp$ g++ -std=c++0x -o tt t.cc && ./tt
9223372036854775808.00000000000000000000 -> 9223372036854775807

Changing this to:
  int64_t v = d < kint64max ? (int64_t)d : kint64max;

/tmp$ ~/build/llvm-opt/bin/clang++ -std=c++11 -o tt t.cc  && ./tt
9223372036854775808.00000000000000000000 -> 9223372036854775807

The question is whether clang or gcc is correct, or there is something
different I'm not understanding going on here. Chandler suggested I file a bug
and let the language lawyers take a shot :)

The AST for the original code from clang looks like this:
     (ImplicitCastExpr 0x31628d0 <col:7, col:27> 'int64_t':'long'
<FloatingToIntegral>
        (ConditionalOperator 0x31628a0 <col:7, col:27> 'double'
          (BinaryOperator 0x31627e0 <col:7, col:11> '_Bool' '<'
            (ImplicitCastExpr 0x3162798 <col:7> 'double' <LValueToRValue>
              (DeclRefExpr 0x3162748 <col:7> 'double' lvalue Var 0x31625e0 'd'
'double'))
            (ImplicitCastExpr 0x31627c8 <col:11> 'double' <IntegralToFloating>
              (ImplicitCastExpr 0x31627b0 <col:11> 'int64_t':'long'
<LValueToRValue>
                (DeclRefExpr 0x3162770 <col:11> 'const int64_t':'const long'
lvalue Var 0x31623d0 'kint64max' 'const int64_t':'const long'))))
          (ImplicitCastExpr 0x3162858 <col:23> 'double' <LValueToRValue>
            (DeclRefExpr 0x3162808 <col:23> 'double' lvalue Var 0x31625e0 'd'
'double'))
          (ImplicitCastExpr 0x3162888 <col:27> 'double' <IntegralToFloating>
            (ImplicitCastExpr 0x3162870 <col:27> 'int64_t':'long'
<LValueToRValue>
              (DeclRefExpr 0x3162830 <col:27> 'const int64_t':'const long'
lvalue Var 0x31623d0 'kint64max' 'const int64_t':'const long')))))"))

-- 
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20130207/f61673b2/attachment.html>