2013/6/5 Rico Schüller <kgbric...@web.de> > On 05.06.2013 08:48, Christian Costa wrote: > >> bone = &This->bones[bone_num]; >> + /* Decrement refcounts of vertices previously influenced */ >> + for (i = 0; i < bone->num_influences; i++) >> + if (bone->vertices[i] < This->num_vertices) /* Ignore out of >> range value */ >> + This->vertices_refcounts[bone-**>vertices[i]]--; >> bone->num_influences = num_influences; >> HeapFree(GetProcessHeap(), 0, bone->vertices); >> HeapFree(GetProcessHeap(), 0, bone->weights); >> bone->vertices = new_vertices; >> bone->weights = new_weights; >> + /* Increment refcounts of vertices newly influenced */ >> + for (i = 0; i < bone->num_influences; i++) >> + { >> + if (bone->vertices[i] < This->num_vertices) /* Ignore out of >> range value */ >> + This->vertices_refcounts[bone-**>vertices[i]]--; >> + } >> > > Please have a look at the code style (brackets {} at the for loops?). It > should be a bit more consistent. Also I think an increment (as said in the > comment) is something like "This->vertices_refcounts[** > bone->vertices[i]]++;". >
Oups! It's the wrong patch. I will send an updated version. Thanks! > > Maybe you could add a test to verify if the comment or the code is correct? > Ok. I'll add a basic skinning test. I need one anyway. Vertices skinning is ok but normals one does not look the same as native. Christian