🚀 go-pugleaf

In <44c6g2$g42@news0.rain.rg.net> Paul Long <plong@perf.com> writes:

>danielv@crt0.crt (Daniel Villeneuve) wrote:
>>
>>From ANSI C 3.5.2.1:
>>
>>``A bit-field shall have a type that is a qualified or unqualified version
>>of one of _int_, _unsigned int_ or _signed int_.''
>>
>BTW, how's this for perverse? The 6.5.2.1 Semantics section says that "A
>bit-field is interpreted as an integral type consisting of the specified number
>of bits." So, as long as we "interpret" the type as integral, does this mean
>that the following is conforming?
>
>   struct {
>      float a : 3;
>   };

Since when is "float" a ``qualified or unqualified version of one of
_int_, _unsigned int_ or _signed int_'' ?

>IOW, it looks like the Standard imposes NO requirements for the type of a
>bit-field. 

It does.  Read again the quote reproduced at the beginning of this
article.

>We just have to interpret the type as integral. Can anybody prove
>different? This would make a lot of C compilers non-conforming, such as acc and
>gcc.

No, it won't.  Your example invokes undefined behaviour, because a 
"shall" requirement which appears outside of a constraint is violated,
so no matter what the compiler does, including crashing and burning,
is the right thing.

Dan
--
Dan Pop
CERN, CN Division
Email: Dan.Pop@mail.cern.ch 
Mail:  CERN - PPE, Bat. 31 R-004, CH-1211 Geneve 23, Switzerland