[Bf-committers] Blender do_versions() patching

Ton Roosendaal ton at blender.org
Sun Feb 24 10:42:46 CET 2008


Hi devs,

I've noticed a quite ugly growth of do_versions() patching in  
readfile.c.
Here's a couple of remarks for everyone who's working there:

1) document the do_versions patch with comments please!

2) (Py)Constraint patching is there double, why?

3) Do not use subsversion increasing, unless absoluteley needed. We're  
now at 15 already, and there's really not that much complex version  
matching needed

4) Most important: design new data in Blender in a way that you don't  
need version patches at all! It's really not that hard... trust Blender  
to initialize new data always as zero, and use that as a initialize  
trigger in your code (apparently, initializing is needed then anyway).

5) If you add code that's not forward compatible, the MINVERSION define  
has to be upped the current number, especially if that'll cause in loss  
of data. I think the particle recode demands it?

-Ton-


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



More information about the Bf-committers mailing list