I'm well aware of what UNIX 03 means. Are you? The core of it is IEEE Std 1003.1. Here's a quote from the Rationale section of the rename(2) page from IEEE Std 1003.1, 2004 edition:
"This rename() function is equivalent for regular files to that defined by the ISO C standard. Its inclusion here expands that definition to include actions on directories and specifies behavior when the new parameter names a file that already exists. That specification requires that the action of the function be atomic."
So to comply with the Single Unix specification, rename(2) must be atomic. The rename(2) function on OS X has been empirically shown to not be atomic. What reasonable conclusions can be drawn here?
The reasonable conclusion is that a bug exists where -- under certain circumstances -- rename(2) fails in a way that is not expected. It's a bug. It's not as if rename(2) is intentionally implemented in a way that is not atomic; it's that there is a bug. Should it be fixed? Yes. Does it mean that OS X is not Unix? No.
One may feel comfortable claiming that OS X is "Unix" based on the fact that all versions of OS X since 10.5 have been "Open Brand UNIX 03 Registered Product" (from an archive of the Apple website), signifying that they have met the requirements for the SUSv3 and POSIX 1003.1 specification.
"This rename() function is equivalent for regular files to that defined by the ISO C standard. Its inclusion here expands that definition to include actions on directories and specifies behavior when the new parameter names a file that already exists. That specification requires that the action of the function be atomic."
So to comply with the Single Unix specification, rename(2) must be atomic. The rename(2) function on OS X has been empirically shown to not be atomic. What reasonable conclusions can be drawn here?