[cfe-commits] Patch for bug #4175

[Chris Lattner]

On Jun 12, 2009, at 10:09 AM, David Majnemer wrote:

> Please understand that this is my first attempt at trying to submit  
> something, I apologize if I have done anything wrong.
>
> Anyway, this patch has some changes to the warnings that we could  
> emit. I personally felt that alternatives would be harder to read.
> If the patch is not up to snuff, I would be more than happy to try  
> to fix it.

Hi David,

This looks nice to me, but it looks like you have some 4-space  
indentation in the patch:

+    if (rType->isFunctionPointerType() || lType- 
 >isFunctionPointerType()) {
+        if (isRelational) {
+          Diag(Loc,  
diag::ext_typecheck_ordered_comparison_of_function_pointers)
+            << lType << rType << lex->getSourceRange() << rex- 
 >getSourceRange();
+        }
+    }
+    if (((!LHSIsNull || isRelational) && LCanPointeeTy- 
 >isVoidType()) !=
+        ((!RHSIsNull || isRelational) && RCanPointeeTy- 
 >isVoidType())) {
+        Diag(Loc, diag::ext_typecheck_comparison_of_distinct_pointers)
+          << lType << rType << lex->getSourceRange() << rex- 
 >getSourceRange();
+    }
      // Simple check: if the pointee types are identical, we're done.

Please change this to indent by 2, include a testcase and I'll apply  
it for you.

Thanks!

-Chris
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20090612/7a699b9e/attachment.html>