New AIX libstdc++ testsuite regression 20_util/allocator_members

[Benjamin Kosnik]

> This is purely an optimization; there's no semantic reason that this
> is necessary. 

Mmmmm. No. Re-read the ChangeLog, and the complete patch as posted on
gcc-patches. Without it, static linking gets duplicate symbols.

There's probably another way to solve the underlying issue, that avoids
breaking AIX: if you could review my original patch on gcc-patches
you'll see what I was up against, and how to see the problem on linux.

I believe the fundamental issue is static linkage of static, non-const,
non-POD data members.

> So, I suggest just #ifdef'ing this on whether or not we're on AIX, if
> that fixes the problem.  It's really an AIX-specific issue -- more
> than, say, a support-for-weak-symbols issue -- in that it's got to do
> with how the names are bound in the linker and such.

Gaaaaah. Please don't. I'd rather add AIX hacks to the testsuite, not
the library.

-benjamin