On Wed, Jun 03, 2015 at 12:54:22AM +0300, Vladimir Zapolskiy wrote: > Of course I'm not a C standard writer, but I have a feeling that one day > bool type may become quite distinct from int (or in other words bool > type will not be one of integer types). This may explain why C99 and C11 > has "_Bool" type, not a hypothetically finalized in future "bool" type. > Also C99 clearly states that macro "false" and "true" may be undefined > and redefined to any arbitrary values. You're confusing several different things here - it's just the same as the handling of NULL and 0. An implementation can make NULL be anything that amuses it (so long as nothing standards conforming can tell) but if you write 0 in a pointer context it has to be a NULL pointer, and if you use a NULL pointer in an integer context it must evaluate to 0. The in memory representation is potentially a separate thing to what the source code does, though practically speaking people are unlikely to implment anything too extravagent. I suspect you'll find that the use of _Bool in current standards is simply to avoid breaking existing standards conforming code that uses bool by suddenly making that a keyword.
Attachment:
signature.asc
Description: Digital signature
