C Semantics in K Framework

[John Regehr <regehr AT cs.utah.edu>]

Hi Derek,

I can't find any language in the C99 standard which forces unnamed 
struct members to be bitfields.  Rather, we see text like this which 
appears to refer to all member objects:

"A structure type describes a sequentially allocated nonempty set of 
member objects (and, in certain circumstances, an incomplete array), 
each of which has an optionally specified name and possibly distinct type."

Can you back up your interpretation?

Thanks,

John




On 8/9/11 1:06 PM, Derek M Jones wrote:
> John,
>
> > The C99 standard seems to make it clear that structs may contain unnamed
> > members.  This argues that KCC and CompCert are correct.
>
> Unnamed BIT-FIELD members.
>
> > Both Clang and GCC (even when given the --std=c99 flag) complain that
> > line 5 does not declare anything, and their output is consistent with
> > this claim.
>
> This is correct.
>
> > Anyone know what's going on?
>
> The code is making use of an extension to C.
>
> > Thanks,
> >
> > John
> >
> >
> >
> > int printf (const char *, ...);
> >
> > struct S0
> > {
> >      int;
> >      int f1;
> > };
> >
> > struct S3
> > {
> >      int f2;
> >      int f3;
> > };
> >
> > union U4
> > {
> >      struct S3 f0;
> >      struct S0 f3;
> > };
> >
> > union U4 g_110;
> >
> > int main (void)
> > {
> >      g_110.f3.f1 = 1;
> >      printf ("%d\n", g_110.f0.f2);
> >      return 0;
> > }
> > _______________________________________________
> > c-semantics mailing list
> > c-semantics AT cs.illinois.edu
> > http://lists.cs.uiuc.edu/mailman/listinfo/c-semantics