I'll keep looking for options. I don't mind sending you the bunch of #include <stdio.h>, <string.h>, <stdlib.h> changes if you don't mind getting them. (I've gone through the whole thing now, and am trying to build from the top again.)
For the include/osgParticle/range file (which shows some evidence of someone dealing with this stuff in the past), I just defined stdlib.h. That isn't the way it is now, so I just wanted to check with whether that's OK. I'll let you decide that when I send it. andy -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Robert Osfield Sent: Monday, June 25, 2007 2:58 PM To: osg users Subject: Re: [osg-users] std namespace on SUNWspro compiler Hi Andy, It sounds like you are having to jump through quite a few hoops, if its STLport causing this then perhaps you could adjust its options. Note, the OSG code is really portable, it works just fine on a wide range of platforms. We've also try experimenting with using cstdlib before but found it just wasn't portable enough, and caused far more problems than it solved. Robert. On 6/25/07, Andy Skinner <[EMAIL PROTECTED]> wrote: > I'd like some guidance on this, Robert. > > I've been going through and adding includes for stdlib.h and/or string.h > as needed, and have gotten pretty far with that. I'm a bit more nervous > as I get to all the directories I haven't been needing to look into > before now (osgDB, osgSim, osgGA). If no one else is getting the errors > I'm getting, then code can continue to be checked in that won't get > stdlib.h explicitly. That might be fine, if I can fix this as I need > them and releases are more frequent now. > > Is anyone else using STLport? I'm not very familiar with it myself, but > I've had some of our infrastructure peopele point to it as a potential > source of the difference between our setup and that of others. > > And a specific question: > In the include/osgParticle/range file, we need to specify that rand() is > std::rand(). There is a using statement commented out. And there is a > section including <cstdlib> if not on SGI, <stdlib.h> otherwise. I > could uncomment the "using std::rand;" statement, or just change the > file to include <stdlib.h> instead of <cstdlib>. I'm not at all sure of > the right thing here. > > Thoughts? > > thanks > andy > > > -----Original Message----- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] On Behalf Of Robert > Osfield > Sent: Saturday, June 23, 2007 6:31 AM > To: osg users > Subject: Re: [osg-users] std namespace on SUNWspro compiler > > Hi Andy, > > On 6/22/07, Andy Skinner <[EMAIL PROTECTED]> wrote: > > One example of a file with these problems is > src/osg/ArgumentParser.cpp, > > complaining about atof(), atoi(), and strncmp(). I don't see stdlib.h > > included specifically, though some std includes are coming in (map, > set, > > iostream, for example). Maybe my versions of those files include > > cstdlib. > > I would have though ArgumentParser.cpp is picking up atof etc from > other OSG headers. > > You could try insider a #include<stdlib.h> at the top of the files > that are having problems to see if that fixes things. > > Robert. > > _______________________________________________ > osg-users mailing list > [email protected] > http://openscenegraph.net/mailman/listinfo/osg-users > http://www.openscenegraph.org/ > _______________________________________________ osg-users mailing list [email protected] http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ _______________________________________________ osg-users mailing list [email protected] http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
