[Verse-dev] pyverse r1p1

[Emil Brink]

On Tue, 3 May 2004, Brecht Van Lommel wrote:

> Hi all,
>
> I've updated pyverse to work with the current verse release (r1p1).
> Support for the Audio functions is still missing. Next to updating, I
> have also changed how pyverse deals with it's namespace. Instead of
> using prefixes like verse_, VN_, V_, ..., it now simply leaves those
> out. So now the python code should look like, for example:
>
> import verse as v
> v.send_node_create(...)
> if type == v.GEOMETRY:
> vmf = v.MFColor()
> ...

Cool! I guess dropping the prefixes makes sense since it's all
imported as "verse" anyway, right?

> I've included the updated tutorial.py, and a very primitive
> wavefront .obj loader example script. The .obj loader hasn't really
> been tested, as I couldn't get connector / loq airou compiled that
> easily with the new verse release, but I thought I'd include it
> anyway.

Cool II! :) I really should get to a point where I can track your
releases and actually test things out from the Python perspective a
little. So many interesting things to do, so little time... Heh.

> It's available here:
> http://projects.blender.org/projects/pyverse/
>
> One thing that seemed odd to me in the verse api is this:
> extern void verse_send_connect_deny(void *address);
> Why is address a void*?

Good question. It looks like it should want to be a string rather
than a random pointer, please comment Eskil. This is something that
could be fixed with just a new patch on release 1 I think, since al-
though it is technically an API change, it's very minimal and does
not change anything in the network. Any sane program would already
pass a string to send_connect_deny and just recompile cleanly.

> Also, verse_send_connect_deny and verse_send_connect_accept pass address
> as an uninitialized variable to the callback function.

Crappy. I'll look into those later, thanks a lot for reporting these
things, it's *very* valuable to us.

Regards,

/Emil