I found and ran the expired 633 for most of the day. The same recipes were used as when I previously crashed. No issues showed up today. But, I think a sampling of just two players may not be good enough. I'd wait for a few others to try before releasing to main.
Agreed… I posted a new puzzle in devprev to test this out one last time (hopefully!)
It's called: ResidueIE Filter Final Test
Please leave any feedback about it here:
http://fold.it/portal/node/993669
Please make sure you are using the latest devprev client and thank you for testing these puzzles!
One would need to run a separate machine with that latest devprev in order not to loose compatibility with the rest of the team running the main client. That may be worth mentioning.
Yes, thank you for pointing this out Bletchley Park.
That is why we want to post the current devprev release to main ASAP!
Hopefully that will be possible tomorrow if there are no issues with the above puzzle.
I fail to understand why it was posted as a non-scoring puzzle. Since this puzzle needs attention of as many different machines as is feasible, it would have seemed to me that posting as a very short ( time-wise ) but scoring puzzle would have more likely have produced a more meaningful result in terms of more machines playing it. I know it won't get the attention of any of my clients in the manner it deserves. Sorry, just my thoughts on the matter.
If we post it as a scoring puzzle, that would require pushing the current devprev client to main, which you recommended against.
The idea of beta testing in devprev is to catch the bugs BEFORE they reach the scoring puzzles.
Otherwise, we end up have to cancel/close scoring puzzles with 0 points, which nobody wants!
I hope this clears up any confusion!
Another machine is not needed, just make a seperate directory for the devprev one.
I would like to see that if the save format is changed again, that first a version that can read both formats and writes the old format is pushed into main. I am not running any tests on running puzzles on devprev because of the format issue and I suspect that that is the case for more people. This results in less testing as normally.
I notice that with the client before this one, in devprev, there still is a small 'handles' leak. It grows slowly. Last I checked there were 90000 handles used after several days 24/7 whereas a restarted machine uses 20000 or so. The number steadily rises when running clients. I have not checked with the latest version as I am very busy with work.
Hopefully this issue is finally resolved for everybody!
Thanks again for your patience, and much thanks to all those who tested this out in devprev to make sure it is safe for main.