On Dec 18, 2007 12:48 AM, Pierre Habouzit <[EMAIL PROTECTED]> wrote: > This is completely insane. "In the past" there was the gcc3 to gcc4 > C++ ABI transition, and that's it.
? gcc3 to gcc4 wasn't a deliberate C++ ABI transition. (I was an active member of gcc upstream at the time.) The major version bump - which shouldn't have happened IMO - reflected the introduction of a whole bunch of new optimizers, no more. You may be thinking of libstdc++5 to libstdc++6, which was a lot longer ago - long enough that I'm not sure when it happened, other than that it was *no later than* gcc 3.0. I don't trust g++ to have maintained ABI compatibility in either direction across releases since libstdc++6, but that's because I'm quite painfully aware of just how easy it is to break C++ ABIs by accident. There's been no *deliberate* forward-compatibility break since then that I know of. As I said to Steve, all the actual breakage I can find evidence for involved gcc3.x, there may not be a problem with the various 4.x series and boost. But I sure wouldn't rely on it. zw -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
