Hi,
Moose policy is to work with the latest of Pharo as much as possible. This implies VM and image.
We still have to get a bit more aligned with the Pharo effort on the image side, but for the VM, we want to stress the latest version of it.
Of course, in exceptional cases, we should make exceptions. In the current case, using stable until the bug is solved is a good solution. But this should be an exception. In the past half a year since we moved closer to Pharo, this is probably the first blocker that took longer than 1 day to get fixed. Thus, the default policy of working with the head works well, and it saves effort in the long run.
Cheers,
Doru