[Bf-committers] IRC meeting minutes, feb 29 and march 7

[Ton Roosendaal]

Hi all,

-------------Feb 29

1. projects overview

Ton points out the public projects listing has been updated:
http://www.blender.org/docs/projects.html
A couple of comments (more python team members) have been added to it.

Two new projects were added and improved; for Mesh tools and Transform

2. 2.33 projects

- Scons
Michel mentions his todo list, when that's done Scons will offer what  
Makefiles has now, plus some more. Jesterking is assisting him on this  
now.
Michel will add 'writing docs' to todo list as well!

- Autoconf
A short disccussion went on around the autoconf topic. Some people have  
mentioned really missing the feature. We could still potentially  
include this in the sources.
Consensus was to 'be firm', and stick to Scons as replacement.
Zr adds that Scons will enable us more things, like integrating python  
controlled test suites, rewriting makesdna in python

- MBall + Mesh
Jiri proposes his latest addition, which allows triangles to be  
rendered using MetaBall Polygonize as well. Three methods for  
integrating this feature in Blender were discussed. One using an Object  
pointer in Mesh pointing to the MBall, two using parent/child for it,  
the third using the current 'family' naming convention. Mainly for  
consistancy reasons Ton cut the knot and decided to use the third  
method.

- New scale option
Theeth describes how this works; while scaling vertices, the  
displacement of vertices will not be proportional to the  
centre-distance, but be constant. That way a sort-of sculpting effect  
can be achieved, where the scaling centre only denotes the direction of  
displacement, but not influences the amount.

3. bugs in tracker

The last 30 minutes of meeting people were diving in the tracker.

-------------- March 7

1. Typo3

The new CMS system is taking off really nice. As of now (march 13) the  
total pages go to 200. Currently these people already have volunteered:

Kent Mein: Python (plugin) section
JesterKing: development section
Ton: releaselogs, BF section
guitarGeek: development pages
_Sysadm: gallery section
Michael Thoenes: education/training section
Bart: overall site admin, did many pages

Goofster, Robertt, Stiv and Michel have accounts for trial too.  
Everyone interested to play with this, or to help managing sections or  
pages just ask me!
Read here more about this cms system:
http://www.blender3d.org/cms/CMS_Manual.97.0.html

2. Game engine

Jesterking currently works on compiling/testing issues. He communicates  
with Kester on merging ODE and Solid further.

Ton agreed to make most of the NaN demos (.blend) online available for  
test and as conformance suite. For gamelogic, the secoac suite can be  
used:
http://download.blender.org/demo/old_demos/

(The testfiles are collected and uploaded, but to make it public I am  
awaiting OK from the artists who made them. Ask me on IRC for a link if  
you need them).

3. 2.33 projects shortlist

- New loopselect
http://www.gvc.net/~jmatthews/blender/loopselect/
Access to these tools are getting a bit confusing. A discussion on how  
to integrate it best didn't give a quick consensus. Agreed was for  
guitarGeek to commit this in bf-blender with his preferred method, and  
review it later on.
(B could be area select tools, Shift-B could be Vert-Edge relation  
select tools
Agreed however was that this doesnt need a specific designed cursor  
type, header-print is sufficient

- New transform
Ton should review Theeths proposal soon!

- DetectiveThorn might pull off adaptive subdivision for displacement

4. 2.33 planning

Aim still is to have this the engine-back release!
For restoration of game engine a couple of tasks are pending:
- compiling Solid for all platforms (Hos will do irix)
- relocate libraries... Solid/qhull should not be in 'extern' but in  
'lib' (right?!)
- verify if we have the latest Solid release

The other topic, python 2.3 support, is pending still. Willian, where  
art thou!

5. Engine post-2.33 strategy

Ton points again to issue with maintenance and developement strategy  
for engine in future. The complexity is huge, and we might not be able  
to provide support at level of other Blender features. We could put  
this in separate project, with own CVS and release cycles. There a new  
team can also redefine functionality and even drop compatibility  
completely.
Hos doubts this would be viable strategy, what would users benefit? And  
why not just use Tuhopuu for that?
This is something the engine coders have to say something about first.

-Ton-

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