>The argument against c++ has been a constantly changing ABI, but with >the release of GCC 3.2 it finally looks like G++ will have a stable API. >time will tell i guess.
i have my doubts about this. with the flexibility that c++ compile-time flags provide, i'm not sure one can ever talk about "a stable ABI" for c++ code, even if the compiler itself never changes it. when used with a dynamic (run-time) linker, things go wrong very quickly unless all the binaries are known to be compiled "identically". --p