A customer opinion on thread protection of static initialization

[Christophe de Dinechin]

We received a message, a couple of days ago, discussing a bug in our  
static initialization scheme with shared libraries. Here is an  
interesting snippet.

| (BTW, don't take this as a criticism of aCC.  In fact, aCC  
generally does a
| very good job on this front.  For example, it's one of the few  
applications I
| know of that generates thread-safe code to guard local static  
initializers.
| That's one reason I'm surprised it has the limitation being  
discussed here.)

I did not make this up :-)
Christophe