
Signature
Steve Maillet
EmbeddedFusion
www.EmbeddedFusion.com
smaillet at EmbeddedFusion dot com
> It's supposed to be 0x500 so if any of the compilers are doing that wrong
> then it's a bug. Easiest way to workaround it is to check for both values.
So it is a bug. The problem with checking both values is that theoretically,
the values might overlap, but unless some clown converts back to using
decimal-based version numbers I think we should be fine. BTW: I found some
entries in the registry that use 500 or something similar for a CE-related
version while other entries seemed to be the binary representation of
L"500". I think that is another of the symptoms or causes of the problem.
All in all, it seems someone made a big mess, starting from the twisted idea
of putting a version pair/triplet into a single number in the first place
and finishing with not documenting it properly. Oh well, someone will work
around it.
Uli