[ODE] Re: Torque: Bug in 0.025, apparently fixed in CVS

[Ruud van Gaal]

...
> i've just checked - you are correct. yes - there WAS a bug in 0.025
> and it WAS intentionally fixed. this is exactly what CVS is good at:
> look at the differences between revision 1.14 and 1.15 in ode.cpp:
>   
> http://q12.org/cgi-bin/cvsweb.cgi/ode/ode/src/ode.cpp.diff?r1=
> 1.14&r2=1.15&f=H

Ah, nice to see where the 0 and the 1 stands for. :)

> the dBodyAddRelForce() and dBodyAddRelTorque() functions were both
> changed to use R instead of R^T. the change log mentions bug fixes.
> v0.025 of ODE uses r1.10 of ode.cpp, so that bug was fixed after that
> release.
> 
> embarrassing. sorry ruud!

That's ok, Russ. It's not called v0.025 for nothing. ;) It's just great that
with the Internet, things like this will be found relatively quickly, and
things get robust much sooner.

> the lesson here is that v0.025 is getting seriously old, don't use it
> anymore. it should not be the default "stable" release that people
> download. i should maybe make a v0.029 from the latest CVS ... ?

Hm, perhaps putting up a small bug list (on your site) for v0.025 might be
quicker. Besides, you may have some bugs in the current code yet which will
then haunt you in much the same way. It's better to spend resources as much
as possible into more productive work instead of administration of things.

Cheers,
Ruud