[PATCH] Destructor support for C++11 thread_local variables

[Roland McGrath]

> On 10/10/2012 05:09 PM, Roland McGrath wrote:
> >> Loosely, construct on first use.
> >>
> >> IOW we need not run constructors on tls variables for a thread until a
> >> function that touches those tls variables runs on that thread.
> > 
> > That's a straightforward specification that's perfectly clear to me.
> > So what's "loose"?
> 
> One must construct all of the tls variables from a translation unit at
> the same time.

Is that a semantic requirement?  Is there a special technical reason for
it, or just "principle of least astonishment" for people who write
constructors with side effects?


Thanks,
Roland