New dlls/ddraw/device.c warnings
After the recent sets of changes, I found that GCC 3.4 issues the following warnings (on FreeBSD 5.4. I checked and depending on how a compiler implements assert(), the warnings are valid, insofar as the compiler doesn't have a way to automatically determine that all code paths are covered: device.c: In function `Thunk_IDirect3DDeviceImpl_2_Begin': device.c:1826: warning: 'FVF' might be used uninitialized in this function device.c: In function `Thunk_IDirect3DDeviceImpl_2_BeginIndexed': device.c:1886: warning: 'FVF' might be used uninitialized in this function device.c: In function `Thunk_IDirect3DDeviceImpl_2_DrawPrimitive': device.c:2752: warning: 'FVF' might be used uninitialized in this function device.c: In function `Thunk_IDirect3DDeviceImpl_2_DrawIndexedPrimitive': device.c:2891: warning: 'FVF' might be used uninitialized in this function Could you have a look into this, Stefan? If someone has a hint on the preferred approach to address this, I can give it a try as well. Thanks, Gerald -- Gerald (Jerry) Pfeifer [EMAIL PROTECTED] http://www.pfeifer.com/gerald/
Re: New dlls/ddraw/device.c warnings
Am Montag 26 Juni 2006 19:13 schrieben Sie: > After the recent sets of changes, I found that GCC 3.4 issues the > following warnings (on FreeBSD 5.4. > > I checked and depending on how a compiler implements assert(), the > warnings are valid, insofar as the compiler doesn't have a way to > automatically determine that all code paths are covered: > > device.c: In function `Thunk_IDirect3DDeviceImpl_2_Begin': > device.c:1826: warning: 'FVF' might be used uninitialized in this > function device.c: In function `Thunk_IDirect3DDeviceImpl_2_BeginIndexed': > device.c:1886: warning: 'FVF' might be used uninitialized in this function > device.c: In function `Thunk_IDirect3DDeviceImpl_2_DrawPrimitive': > device.c:2752: warning: 'FVF' might be used uninitialized in this function > device.c: In function `Thunk_IDirect3DDeviceImpl_2_DrawIndexedPrimitive': > device.c:2891: warning: 'FVF' might be used uninitialized in this function > > Could you have a look into this, Stefan? If someone has a hint on > the preferred approach to address this, I can give it a try as well. This seems to be a clear case of 'stupid compiler'. The assert(0) in the default case of the switch statement will terminate the app, but it the compiler thinks that the call below will be called. I sent a fix which replaces the assertions by an ERR log and an error return to wine-patches Thanks for the hint Stefan pgpgWt1poompr.pgp Description: PGP signature
