devkitPro

I'm a bit confused by vertex formats and vertex descriptions in libogc, for example: So GX_SetVtxAttrFmt takes a GX_VTXFMT0 format argument, but GX_SetVtxDesc doesn't. I looked through the code in libogc, and it always writes the vertex description into CP registers 0x50/0x60.
If you look into YAGCD, there should be one vertex description for each vertex format. Or is that Wii-only hardware? Otherwise I don't see the point in being able to declare several formats, if the actual declaration has to consist of the same elements each time. Or are you expected to send a new declaration whenever you render primitives with a different format?

Hi,

you determine the VtxAttrFmt to be used on call to GX_Begin. That's why you don't have to set this with a call to VtxDesc explicitely.
This way you can setup all your used VtxAttrFmt(s) at render init and just select them by using the appropriate fmt slot on call to GX_Begin.
SetVtxDesc is more or less only used to specify what vertex data will be in the stream following GX_Begin.

regards
shagkur

The reason I was asking is, the VtxDesc contains which elements of a vertex are present, and if they're direct, or if they use an 8 or 16 bit index. So if you've got different shaders it would be something you're changing very frequently. On the other hand I can see myself settling for only a handful of VtxAttrFmt. But I guess if that's how the hardware is designed I'll have to work with that.

Hi,

reading up the patent again says clearly there's only one vertex descriptor called "current vertex descriptor". So in my opinion there's a mistake in YAGCD at this point.