[Henri Verbeet]

On 25 June 2010 18:07, Misha Koshelev <misha680 at gmail.com> wrote:
> So here is why I thought maybe having a separate shape.c file might be good. I am very new to D3DX9 so
> please pardon my ignorance, but simply looking at SDK, I see:
>
> i) http://msdn.microsoft.com/en-us/library/bb172976%28v=VS.85%29.aspx
> Shape Drawing Functions are all defined in "D3dx9shape.h"
>
> whereas
> http://msdn.microsoft.com/en-us/library/bb172973%28v=VS.85%29.aspx
> Mesh Functions are all defined in "D3dx9mesh.h"
>
> This is not a reason to implement shape drawing functions in a separate file _per se_, but additionally:
>
> ii) all the tests in mesh.c do not rely on CreateWindow calls or Direct3DCreate9 calls whereas those for shape
> functions inherently will (like line.c)
>
> Thus, since I'm going to try to focus on d3dx9 stuff (hope code-freeze is over July 4th), it seems like it might be good
> to have a separate file.
>
> What is the argument for keeping them all in the mesh file besides that they are mesh creation functions?
>
Essentially that there's no convincing reason not to. It's only a
handful of functions, clearly related to the mesh functions, so you
can't justify it with reasons like prohibitive source file size or
being clearly distinct functionality. Also, Wine as a project has a
preference of fewer large files over lots of small ones. You can make
good arguments for that in terms of e.g. ease of editing and keeping a
good overview of things, but if nothing else you can simply consider
it project policy.