Yeah, the confusion with the memory bug was partly due to me. After your original fix that still had a problem with the reader I introduced the bug again when I tried to fix that.
I agree that your latest fix should definitely work. In fact I took the opertunity to install anteater and whereas before I could run the test script for this bug 4300 times before memory creeped up to 115MB I am now at iteration number 5500 and seeing a stable 50MB's. I haven't yet voted, but +1 for releasing! Unico Carsten Ziegeler wrote: > > To increase the confusion :) I think we should stick to the > code freeze as I see the chance for a release tomorrow. > > As far as I see it, we have the vetos from Reinhard and Joerg > (Vadim revoked), and it seems that these vetos stand and fall > with the memory leak bug, so all we have to do is: fix the > bug, revoke the vetos, do the release. > > I just committed a fix for the bug which imho should work. So > we only need one to cross-check this. > > (I really start to like calling out "code freeze", "end of > code freeze" and repeating this over and over again) > > Carsten > > >