[ocaml-opengl] LablGL's Raw module

Jacques Garrigue garrigue at math.nagoya-u.ac.jp
Wed Mar 13 22:55:05 GMT 2013


On 2013/03/14, at 4:15, Florent Monnier <monnier.florent at gmail.com> wrote:

> Hi,
> 
> Does someone knows why the module Raw has been written for LablGL
> instead of using Bigarray's?
> 
> Is it just that bigarrays didn't exist yet when LablGL was first
> written? Or is it for a better integration with the module API?

Originally, this is just that bigarrays didn't exist yet.
Then I thought of moving to bigarrays, but there wasn't a 1-to-1
correspondence, so it finally never got done…
Maybe it's not too late to do the jump.

LablGL has been around for 15 years, and it really shows…

> IMHO it makes interoperability with other components, other libs more
> difficult (but maybe the solution would only to externilise Raw from
> LablGL) both related to the type system dependency, and/or we have to
> copy to and from this data structure instead of just using ba's
> everywhere while ba's are available from the base standard library.

Concerning interoperability, Lablgtk somehow overcomes this with
the Gpointer.region abstraction, which just describes a memory
area, and can be created from any such structure.


Back to LablGL, I'm really sorry I cannot give more time to it.
If you look at the git repository, you will see that there is a new
multitex branch by Didier Cassirame, with a lot of useful
additions.
If you have some comments about it, let me know.

Jacques Garrigue


More information about the OpenGL mailing list