Re: Thoughts on CPU set manipulation functions

[Kalle Olavi Niemitalo <kon at iki dot fi>]

"Bert Wesarg" <bert.wesarg@googlemail.com> writes:

> Sure you can get the result by:
>
>   cpu_set_t tmp;
>   CPU_COMPLEMENT(tmp, s2);
>
> Oops, there is no CPU_COMPLEMENT()!
>
> Hmm, ok we have an CPU_XOR(), unlucky us we don't have something to
> set all bits (ie CPU_ONES())...
>
> Hmm, sorry I'm out of options.

s1 & ~s2 is the same as s1 & (s1 ^ s2), or s1 ^ (s1 & s2).

I don't have the documentation for CPU_AND so I don't know
whether one is allowed to specify the same cpu_set_t object for
input and output.  If not, this would require a temporary.  But
anyway, the &~ operation can be implemented with the current
interface.

Attachment: pgp00000.pgp
Description: PGP signature