From: Roland (roland.schwarz_at_[hidden])
Date: 2003-12-18 13:37:15
(Michael Glassford <glassfordm_at_[hidden]>) wrote:
> Michael Glassford wrote:
>
> > Roland wrote:
> >
> >> I found myself, posting a suspected bug in the (Windows)TLS code...
> >
> > What suspected bug in the Windows TLS code did you report? I don't
> > remember it, unless you mean that you consider not being able to link
> > the Boost.Thread library a bug.
> >
>
> I never noticed an answer to this question. What was the TLS bug you
> reported?
>
Oh, this might be due to my unfortunate quoting style. My message (the very same
upon you were asking this question) contained a copy of the original bug description.
I requote the relevant part:
"Problem: When using release on a tss variable, the associated cleanup routine seems not to
be updated correctly as it tries to delete the original pointer, instead of the 0-pointer."
I accompanied this with a small example that displays the error. I am assuming, that calling
release should relinquish control of the contained pointer, but in fact it does not, which results
in doubly freeing the data.
Roland