There should be a different lint define for this case, rather than allowing set waitfor when people expect no sleeping.

We are not the only people using this so we can't change the API contract like that.

But set waitfor doesn't sleep!

It just guarantees the caller will not be locked up, the same thing should_not_sleep tries, but fails to guarantee.

the only time set waitfor = false would kick in, is in a sleep that should_not_sleep doesn't catch.

And because byond is not a statically typed language, should_not_sleep can not catch every case where something can sleep, trying to is as mathematically impossible as the halting problem, you can get very close, but not perfect detection.

It makes no sense to force disabling of set waitfor = false to use should_not_sleep, that seems like the kind of thing where, "We are not the only people using this", so we shouldn't be telling them if they can or can not use both at once.