[boost] Re: Thread-Local Storage (TLS) and templates

From: Alexander Terekhov (terekhov_at_[hidden])
Date: 2003-02-19 12:55:38

Peter Dimov wrote:
[...]
> I must admit that I didn't expect even for p to be different across threads.

Not doing so would probably "violate" a rather nice axiom stated 
in the 1.7/1. ``Every byte has a unique address.'' ;-) Seriously, 
TLS shall not be "private"; similar to automatic storage duration, 
an object with thread-local storage duration shall be accessible 
by all threads... after the *owner* thread has "published" its 
address. And, BTW, lazy-vs-eager allocation/creation is another 
rather interesting issue...

regards,
alexander.