[cfe-commits] [Patch 4 of 5] Proposed fix for PR12850

[Andy Gibbs]

On Saturday, August 25, 2012 7:34 AM, Richard Smith wrote:

    Could you give me some example code for your "nasty corner case", and I'll
    fix my code.  Thanks.



  extern const int n;
  constexpr int get() { return n; }
  constexpr int n = get() + 1;
  constexpr int m = get();
  static_assert(m == n, "");

Wow, that is nasty and my first thought was: is that even legal?

This is the output from gcc 4.7.1:

~/testcase$ g++ -std=c++11 horrible.cpp 
horrible.cpp: In function 'constexpr int get()':
horrible.cpp:2:33: error: the value of 'n' is not usable in a constant expression
horrible.cpp:1:18: note: 'n' was not initialized with a constant expression
horrible.cpp: At global scope:
horrible.cpp:3:15: error: redeclaration 'n' differs in 'constexpr'
horrible.cpp:1:18: error: from previous declaration 'n'
horrible.cpp:4:23: error: 'constexpr int get()' called in a constant expression

What does the standard say? (I haven't got a copy myself.)

Cheers
Andy

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20120825/0439f696/attachment.html>