Static data in dynamically loaded C++ shared libs on linux

[Martin v. Loewis]

> It is not a nonconformance with the ANSI standard for C++, but it is a
> bug.  I'm not into this, well we know it is stupid and wrong, but it
> isn't a bug thing.  Let's call it what it is, a bug.

Calling it a bug implies that it can and ought to be fixed everywhere
it occurs, without breaking anything else. I'm not so sure this is
possible.

> There are solutions that can be implemented to solve this bug.  One of
> them is an atexit with two parameters.

Unfortunately, there is no such atexit routine, and we cannot provide
one, either - it must be provided by the C library, which we don't
provide.

> See HJ Lu's (or was it Drepper?) excellent previous posting.

If you are referring to an earlier patch in this area, well, this was
actually mine...