PR 13583

[Paolo Carlini]

Loren James Rittle wrote:

> I think this race is still present but on most/all CPU
> arches, it is a memory leak at worst. It is bounded
> by number of threads... I suggest we mark it as will
> not fix for 3.4 release.
>
> Comments?

Hi Loren. Well, I agree with your analysis, and vote for NONFIX
3.4. However, I believe that for the next releases we really
should work more systematically and consistently on the various
MT related issues...

Paolo.