1008 segfaults in genattr

[Robert Lipe]

> I refuse to accept disabling HAVE_ATEXIT until someone can explain why it's
> a problem.  So far the explanations haven't cut it.

OK.   Between the explanations in this thread and the ones under the 
titles of related unixware and openserver problems, I don't know any
better how to better explain it.   Yes, perhaps I'm still missing the
Real Problem, but losing HAVE_ATEXIT seems to make life better for some
cases.   It seems to make life better for a Lot cases under Unixware. 
If we're masking some other problem, and you refuse to accept this, 
we're going to have to roll up our sleeves and understand it.

What do you need?   A disassembly of binary compiled with a 
"HAVE_ATEXIT" compiler that fails p2732b.C?   

I've included stack backtraces, disassemblies, and objdump output
in previous threads.

I can't provide this for Unixware, only OpenServer.

RJL