we have a similar, but not quite the same, routine in wined3d too, I think in device.c. It creates a wined3d vertex declaration from a FVF, while the d3d9 one creates a d3d9 declaration. Can you see if the wined3d routine is affected too?
I didn't unify the wined3d and d3d9 code because we'd need need some wined3d->d3d9 decl backwards conversion which is probably as bug prone and maintainance heavy as an almost-clone of the FVF->decl conversion routine, and I dislike the idea of creating a d3d9 decl in wined3d. If anyone has a nice solution for this feel free to suggest it. > -----Original Message----- > From: wine-patches-boun...@winehq.org [mailto:wine-patches- > boun...@winehq.org] On Behalf Of Chris Robinson > Sent: Sunday, December 28, 2008 6:59 PM > To: wine-patc...@winehq.org > Subject: d3d9: Use the correct size for D3DFVF_XYZW [try 2] > > Now with a test.