[Bf-committers] New "merge cdlayers by name" feature can generate more than 8 allowed uvmaps or vcol layers

[Brecht Van Lommel]

The only reason there is a limit is indeed for blender internal, where
it's just these two lines in struct ShadeInput:

    ShadeInputUV uv[8];   /* 8 = MAX_MTFACE */
    ShadeInputCol col[8]; /* 8 = MAX_MCOL */

I see two good solutions:
* Support dynamic allocation of these things. It needs to be done
carefully with per-thread memory pools in the Render struct for
performance, and would need a shadeinput_free function because all of
ShadeInput is just static now.
* Document this as a limitation of Blender Internal and allow
unlimited layers elsewhere. Even if the mesh has that many UV layers,
it's unlikely that one material would use that many?

Brecht.


On Mon, Nov 11, 2013 at 9:55 PM, Bastien Montagne <montagne29 at wanadoo.fr> wrote:
> When you merge two meshes with each 8 UVMaps which names are not
> matching, you get a mesh with 16 UVMaps (same goes for vcol layers).
>
> This is not completely bad per se, as the 8 limitation is only at the
> shading (render) level. However, we should try to fix it - here are some
> solutions I can see:
>
> 1) Let code as it is currently, warning the user when too much layers
> get generated, and add tools to merge layers (in a same object).
> 2) Just drop the layers above limit number (easy to implement, but not
> user-friendly at all!).
> 3) Fallback to "non-matching-name" merging when we do not have enough
> room to create all needed layers (not really good solution either, imho,
> as it would give some rather unpredictable results on user POV).
>
> So I would rather go for 1)… We would probably also need some API from
> customdata.c to easily check the max number of layer is respected, and
> optionally merge or drop the supernumerary ones.
>
> Any feedback is welcomed here, as usual, especially if you have better
> ideas on how to handle this issue!
>
> Best regards,
> Bastien
> _______________________________________________
> Bf-committers mailing list
> Bf-committers at blender.org
> http://lists.blender.org/mailman/listinfo/bf-committers