Bruno Haible wrote:
How about adding a member #if @HAVE_LONG_LONG_INT@ long long int __li; #endif to this union? Not that it would make a difference on the existing architectures, but it may be more future-proof.
I thought about that, but my kneejerk reaction was that the likelihood that a new architecture would come in without C11 support was low enough (and declining) that it didn't seem worth the trouble of adding a dependency on the longlong module.
But perhaps I'm underestimating how long Microsoft will take -- is that the main worry here? If so, how about throwing ptrdiff_t and size_t into the mix instead? That should catch the 64-bit Microsoft Windows platforms, without adding a dependency.
