On Wed, Sep 16, 2009 at 3:35 AM, Maxim Dounin <mdounin@mdounin.ru> wrote:
> Patch looks correct for me.  It looks a bit fragile though,
> probably we need a bit more bulletproof code here.

Would you mind elaborating on what you think is fragile about it?
E.g., how would you rather this bug be patched?

> More generally - resolver known to leak, and probably requires
> code audit.  It would be fine if you look into it.  I believe
> Artem Bokhan will help with testing (cc'd as I'm not sure he is on
> English list).

Great. I'll spend some time looking at the rest of the code then.

Thanks.