Thread safety and static initializers

Thu Oct 11 01:14:00 GMT 2001

Looks like I was wrong about the static initializers problem.
However:
Bryce McKinlay writes:
 > Anthony Green wrote:
 > 
 > >Am I the only one who would trade the static initializer bug for the alleged
 > >thread safety bug?
 > 
 > Maybe there isn't a thread safety bug anyway. As long as all our dlopen 
 > calls are synchronized (maybe dlopen is thread safe anyway?) then I 
 > don't see a problem.
Ah, right. I never did get to figure out what the supposed thread
safety problem was.
It sounds to me like this could simply be fixed by either
a. Showing that dlopen() is thread safe -- Uli could tell us this.
or
b. Wrapping dlopen() with synchgronization code.
If that's all there is to do, I'll do it straight away!
Andrew.