> My concern is specifically about options for changing the default > language version, not options for changing the libstdc++ ABI. More > generally, about configure options affecting the semantics of code > passed to GCC, as opposed to non-semantic configure options such as > choosing the processor for which to optimize. (The closest we have > to that at present is the options controlling decimal floating point > and fixed point, but in general those are enabled by default wherever > they work, and they only serve to permit some programs that would > otherwise be rejected.)
Thanks for the clarification. I understand your concern here, and frankly it seems appropriate. I'm not quite sure what is going to work best here... Let me re-think this and try to come up with actual patches that do something sane, so that I can see the true scope of this work. At that point (stage 1) I'll come back with more detail. best, -benjamin
