[Verse-dev] R5?

[Emil Brink]

Eskil Steenberg wrote:
> Hi
> 
> 
>>Wasn't  it decided to release the the code a looong time ago?
>>Are we waiting for some specific features or bugs?
> 
> 
> My bad, I asked emil to hold back on the release, there was one thing i
> wanted to get in there its a change in the way ids are assigned for
> material fragments, vertices and polygons.
> 
> I have just written up the patch, and i assume emil will put it in CVS as
> soon as he grats back from the festivale.

Okay, done. Not tested the new code at all, but existing programs (did
a quick upload of a bunny and viewed it with Quel) seemed to work.

> I think i have described the change before, but the idea is that thehat
> when it comes to thees ides you can now either choose you own ids (of you
> choose low ids) or you can send them as -1 to let the server assign the
> ids.

Yeah. I should update the spec now, I guess.

> This souldent break any existing apps but will help if you write geometry
> code and you dont want to implement a find-empty-vertex/poly-id function,
> and it also helps alot if you what to uploade more then one fragemnt at
> the time and make them link in to trees.
> 
> I have no more featiures i want exept one: it should work. Right now
> people are experiencing encryption bugs where verse claims to have goten
> borken pakets. This should be fixed before a releas is made. I will work
> with emil on this as soon as he gets back.

True. :/ It's really annoying, those crypto problems. A related problem 
is that we've had feedback from a guy at Nada's theory group about 
better crypto code, but we (or at least I) don't know his name. :/ There 
was talk about him having code for AES encryption, to replace the crap
XOR we use now, but since we can't contact him ... :/

> I would also like get my performance patch running, but it doesnt break
> compatibility so we can do it as a fix after the release just as well.

It breaks stuff for me, at least, so it needs looking into.

/Emil