From: Joschka B. <jos...@am...> - 2008-03-24 14:55:28
|
Hi all, On Mar 24, 2008, at 3:19 PM, Yuan Xu wrote: > I updated the new preview of rcssserver3d-0.5.7 in > http://xuyuan.cn.googlepages.com/rcssserver3d-0.5.7-preview.tar.gz > Thanks for putting together the release preview, Yuan! > Could you take a look at it? > if there is not any problem, I will publish it tomorrow. I tested the code and came across a few issues, but I think they're all Mac OS X (Leopard) related: * I had to add an include to lib/oxygen/simulationserver/ netmessage.cpp to provide otherwise missing types (I didn't have the problem in Tiger). * I added the directory 'macosX' to EXTRA_DIST in the top-most Makefile.am; I couldn't compile the code without the files in that directory * I had conflicting typedefs in /usr/local/include/ode/config.h and a system header (again, this was no problem in Tiger). I didn't know an elegant workaround for the problem right away, so I had to comment out the 'typedef unsigned short uint32' in the ODE config header; this way, the code compiles, but this hack is -- of course -- not satisfactory :-( Does anybody have a good idea how to prevent this conflict? * I noticed that the user's manual is missing the references; maybe it was created without running Bibtex? Also, there are still some things missing and incorrect in the manual (I know for sure that the table with dimensions of the Soccerbot is still wrong, I didn't have time to correct it yet :-( ), so we should maybe tell people that this is still work in progress (but feedback is welcome of course :-) ). After successful compilation, I could run the server without problems, but when I connected agentspark (using the composite bodies), I got the same error as reported by Feng Xue on the sserver-three-d list today (using ODE 0.9): ODE INTERNAL ERROR 2: The centre of mass must be at the origin. in dBodySetMass() (Core) caught signal 6 (Core) exit Does anybody know what could cause this? I had this running without problems before. Cheers, Joschka |