> > Overflow does *compile* with 2.96, but not everything works as it's supposed to > > (don't know yet whether the problem is my code or gcc 2.96) > > Just an academic question here: Is the problem that we're having with GCC 2.96 > related to C++ standards changing? Again, I'm not a C/C++ expert, but I'm a > little surprised that a "minor" upgrade in GCC can break so many things. I'm not too sure about the problem... it may actually be related to libstdc++, but I'm not sure... BTW, there's a lot of differences between 2.95.x and 2.96 (which is *supposed* to be experimental stuff leading towards 3.0). Actually, I've even found that if I try compiling with -O0, the link fails because of internal g++ symbols, so that's at least a bug... Anyway, the result from the limited time I tried compiling/running on a RH7 box is that wierd things happen. I didn't investigate enough to tell whose fault it is. I'm waiting for gcc 3.0 to come out and make sure that it can compile Overflow correctly. Right now, I thing the only problem with gcc 3.0 is some "fludge" I use in the which is not ANSI C++ and which gcc 3.0 screws up. Jean-Marc