Another thought on this. Most problems seem to surface as one of two symptoms:
- People putting their laptop into a backpack (lid closed) where it overheats after inadvertently waking up, or
- People leaving their laptop unplugged after suspending it, and it runs out of battery after inadvertently waking up.
What if we simplified the whole scheme and always put the laptop back to sleep (after any system wake) if all of the following are true:
- The lid is closed.
- No external monitor is connected.
- AC power is not plugged in.
I can't think of any situation where this would cause undesirable regressions, and it would completely fix symptom (1) as well as partially fix symptom (2).
As for symptom (2), the situation is not the worst if AC power is plugged in - the laptop will consume some power and perhaps interrupt a person's sleep at night with a bright display backlight, but the battery won't die. If power is disconnected when the laptop comes out of sleep, and the user did not close their lid or leave an external monitor connected, then we still have an issue. But no regression compared to before.
Regression-free and relatively simple feels like a solid approach to me, and perhaps we can still dig into ACPI and wake-up reasons in the future as a complementary improvement. What do you think?