Il 07/10/2011 03:36, Eli Friedman ha scritto: > Atomic ops; similar to the patch I submitted earlier, but this version > adds the restriction that the first operand to the __atomic_* > operations must be an _Atomic(T)*. I'm sending this to double-check > that I haven't missed any serious issues. > > After this is committed, I'll put together a patch for > __atomic_is_lock_free, including the necessary machinery to come up > with the right answers.
Can you explain why these builtin need to have another node (AtomicExpr) and cannot be handled as ordinary CallExpr like other builtins? _______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
