[Bf-committers] Meeting minutes, sunday december 12, 2005

Ton Roosendaal ton at blender.org
Sun Dec 12 18:06:47 CET 2004


Hi all,

1) Main topic; release 2.36 or not!

Although most agree it's about time to do, the status of the engine  
deserves to be at least on 2.35 or 2.35a or 2.36-RC1... currently also  
the 'regression' files crash for it (conveyor belt).

To not frustrate Kester too much, I hope he can find the quick solution  
for the current error(s). Mal has disected the tracker #2006 bug, and  
posted information there. I have cvs updated older versions to verify  
when it's wrong. And apparently it's in the commit(s) done december 5.  
Or more exact;

Doing a:
cvs update -D "December 4, 2004"
in source/gameengine/

Gives an engine surviving bug #2006 and the conveyor belt. A cvs of one  
day later crashes.

Proposal as discussed and agreed on is:

- Wait for Kester to find a quick solution or a fix
- When this can be in the next 24 hours, we move to release modus
- Othwerwise, since tuesday not all release managers have time,  
wednesday is earliest for another release attempt.

Of the release builders, Yann (Linux), Jean-Luc (OSX), Chris (Irix),  
Simon (Windows) were in the meeting, and they're all standby.

Will keep everyone updated here!

2) Other projects

About new projects, please keep checking wiki, especially for updates  
on the Dependency Graph (Jean Luc) and Softbody (Jens Ole) projects.

Also briefly was discussed the future of the engine in general. As  
discussed before, it could be a good option to provide the engine as an  
optional plugin (like yafray). With as benefits;

- engine and player can get own release cycles
- invite 3rd party engines to check on integrating with same API
- if API is well written, it's easier to switch engine plugins for  
doing more radical changes, like supporting OpenGL pixel shaders (could  
be time for a redesign there too)
- we can bring back in Blender the simple 'Enji' engine (old 2.04  
engine), which isn't as advanced, but also a lot easier to maintain.  
Potential GUI conflicts can make it frustrating though... but best part  
of Enji is that it re-uses all of Blender internals.
(Since I coded this engine; I would redesign Enji from scratch too...  
but that's a load easier work than doing it for current engine :)

Disadvantages;
- it separates teams...
- this way engine loses status, lots of potential users (needs extra  
install) and loses the pressure of being part of Blender release cycles

Still, a possibility worth checking on. For the next months...

-Ton-

------------------------------------------------------------------------ 
--
Ton Roosendaal  Blender Foundation ton at blender.org  
http://www.blender.org



More information about the Bf-committers mailing list