[PATCH] C11 _Bool bitfield diagnostic

[Hubert Tong]

The values of _Bool is not limited to 0 and 1 in C (although storing such a value is not straightforward). It is conceivable that some platforms would have _Bool, aside from conversions, behave the same as unsigned char. My earlier statement that the message is sufficiently generic does not take this consideration into account.

It seems making this a plain warning specifically for _Bool on a portability basis is viable; however, the best argument I could come up with for why _Bool warrants special treatment is that it is somewhat likely to have padding bits. My guess is that implementations which consider _Bool to have padding bits are likely to consider all cases where the padding bits are non-zero to be trap representations.


http://reviews.llvm.org/D10018

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