[PATCH] [Sema] Don't warn if taking the absolute value of a difference of unsigned values.

[Richard Smith]

The difference of two unsigned promoted types is also unsigned, so cannot be negative. Perhaps we should word the warning differently in this case, but a warning does seem justified.


http://reviews.llvm.org/D10263

EMAIL PREFERENCES
  http://reviews.llvm.org/settings/panel/emailpreferences/