"Michael Glassford" <glassfordm@hotmail.com> wrote in message news:ce8eev$glv$1@sea.gmane.org...
Roland wrote:
On Wed, 28 Jul 2004 11:06:39 +0200 Bronek Kozicki <brok@rubikon.pl> wrote:
[snip]
I wasn't either. It looks very interesting, but what Roland points out below is a definite problem.
But as I can see (after a quick look) the callback will not be scheduled to run in the context of the exiting thread. But this is a requirement to be able to free TLS. This is different to THREAD_DETACH which still is scheduled into the exiting thread. This might also be the reason why one should call TlsFree from there.
Either I missed something (jumped into the discussion at about this point) or just don't understand, but why should TlsFree have to be called in the context of an exiting thread? // Johan