On 7/18/23 14:57, Haitao Huang wrote: > Okay, that explains. I would consider it still triggered by high > pressure blips 😄 I'm talking about a "blip" being a single allocation. It's *LITERALLY* the smallest (aka. lowest) possible quantum of memory pressure. So go ahead and try to write the changelog without "high" or "low". But, sheesh, if you and are somehow using "high" and "low" to describe the exact same condition, I think that's a rather large communication or understanding problem somewhere. It doesn't bode well for this simple patch.
