Physics system in the Blender Game Engine
Oct 13 2017
I can confirm the problem.
If I try inversed action to copy pivot offset array of rigid body joint constraint to another location array, I have an error message.
Apr 20 2017
Apr 7 2017
Feb 13 2017
Sep 2 2016
Apparently I had an unsubmitted comment still laying around. I haven't done anything with the GE since the start of the year, so I trust @Mitchell Stokes (moguri)'s opinion of this patch.
Aug 22 2016
This is a simple file showing the creation of a new mesh and the proper replacement of both the graphics and physics mesh.
Overall, this patch looks good to me. I like that it extends the API, which will help keep compatibility (as opposed to making changes deep in Physics code). As long as CcdShapeConstructionInfo's copy constructor was not getting called before, I do not see how this can break existing games.
Aug 16 2016
Apr 9 2016
Can I get an update on what's still missing?
Mar 10 2016
It's no more working, i can't find a solution...
Jan 25 2016
This is discussed in D1545 too.
Jan 22 2016
Jan 21 2016
First bad commit:
Author: Sybren A. Stüvel <email@example.com>
Date: Fri Oct 9 12:12:13 2015 +0200
What web-browser / OS?
Jan 19 2016
Excuse me, When I try to upload files, goes to unknown page and fail.
Please always add files here…
Dec 21 2015
Added some comments documenting the additions.
The already previously made changes applied again.
I updated everything so the diff is relative to master so hopefully changing the diff is easier now.
Ok so my local git diff shows something different than the diff I have here, not really related changes I think, it's just a lot longer. What can I do about that? I mean if it's unrelated you'll probably not want it in the patch.
Dec 16 2015
Fair point I'll get to it on the weekend.
Once I fixed the compiler error, the patch seems to work. There are a few things missing, though: clear documentation. I know that the Blender code isn't the end-all-be-all of well-documented code, but IMO we should at least properly document new stuff.
Dec 15 2015
Please update the summary to give us some more information. Right now, you don't describe any problem with the existing code. You just explain what the new code does, but not why.
I updated the suggested changes (finally).
Dec 2 2015
This is caused by D1545, which allows setting the velocity to zero. Until we fix this, you can work around this by enabling "Add" for the linear velocity. That way it doesn't set it to 0, but just adds 0 and doesn't influence the velocity.
Dec 1 2015
Nov 20 2015
Tested and works fine for me too (debian64 testing, NVidia 850m). Will archive for now.