Re: [PATCH 1/4] wined3d: Properly up-scale WINED3DFMT_R5G5_SNORM_L6_UNORM.

[Stefan Dösinger]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Am 2015-03-03 um 12:18 schrieb Matteo Bruni:
> AFAIK C doesn't give any guarantee on a specific allocation of 
> bitfields in memory. I'm sure this works in practice with GCC and 
> maybe also with the other compilers we care about, but if it
> doesn't get too ugly I'd prefer explicit shifts and masks and be
> safe.

My C / C++ book agrees with you. The compiler is free to add padding.
We can force the compiler to add padding, but not prevent it from
doing so. As much as I liked the idea of using a struct here it won't
reliably work.

C++ adds a few more features to bitfields, but we don't use C++.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2
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=eoL1
-----END PGP SIGNATURE-----