Re: [boost] Enhanced call_once()

From: William E. Kempf (williamkempf_at_[hidden])
Date: 2002-08-06 11:46:07

----- Original Message -----
From: "Pete Becker" <petebecker_at_[hidden]>
To: <boost_at_[hidden]>; <boost_at_[hidden]>
Sent: Tuesday, August 06, 2002 11:19 AM
Subject: Re: [boost] Enhanced call_once()

> At 10:36 AM 8/6/2002 -0500, William E. Kempf wrote:
> >I've discussed this with Mr. Becker before, and since he dropped the
thread
> >I assumed we agreed it wasn't possible to implement his interface. Now I
> >guess this isn't true.
>
> It never was true. I dropped the thread  because you made some conclusory
> statements about what I had in mind, although I had said no such thing. I
> saw no point in continuing.

Sorry, you should have said so.  I'm not sure what conclusory statements you
think I made, but I probably was only trying to understand your thinking.

> >True static initialization means that it occurs at compile time, not at
run
> >time.  This eliminates the initialization order problem, since any needs
for
> >runtime thread synchronization are absent when initialized at compile
time.
> >This type of static initialization is possible in C++, but requires POD
> >types.  The design Mr. Becker proposes doesn't use a POD type, and thus
> >construction will occur at run time, and this means run time
synchronization
> >is required.
>
> Construction at runtime does not require runtime synchronization. You're
> making some assumptions here, and I don't want to guess what you have in
mind.

When other constructors can start threads (and I don't think we can or
should preclude this) then initialization order issues mandate that
construction at runtime requires runtime synchronization.  Again, I'm very
interested in a solution that works and follows the design you've suggested,
so if it can be done I'd very much appreciate hearing how.

Bill Kempf