On 7/4/22 20:30, Tom Tromey wrote: >>>>>> "Tom" == Tom de Vries writes: > Tom> Fix this by guarding the write with a lock. > > I would rather we avoid locks. Ideally the existing exclusion mechanism > should be made to work, but if that fails, perhaps we can use another > atomic. I gave a different approach a try: eliminating the field. That seems to work reasonably well, apart from one specific usage that I disabled (mentioned in commit message). I'm not sure this is a good idea though, but I thought I could at least post it. Currently testing. Thanks, - Tom