After looking through the code, Pat's thoughts in comment 14 seem right to me.
For indicator-datetime's part, it can pass the system bus into the object constructor instead of making an asynchronous call to fetch it. This narrows the dangerous interval in datetime between wakeup and calling requestSysState -- though, NB, I'm not able to reproduce this bug in testing and so I'm not able to determine how much this change ameliorates it.
Outside of indicator-datetime, this might be a question for repowerd wrt ensuring that, after system wakeup, apps have a chance to respond to the wakeup before the system goes back to sleep.
After looking through the code, Pat's thoughts in comment 14 seem right to me.
For indicator- datetime' s part, it can pass the system bus into the object constructor instead of making an asynchronous call to fetch it. This narrows the dangerous interval in datetime between wakeup and calling requestSysState -- though, NB, I'm not able to reproduce this bug in testing and so I'm not able to determine how much this change ameliorates it.
Outside of indicator-datetime, this might be a question for repowerd wrt ensuring that, after system wakeup, apps have a chance to respond to the wakeup before the system goes back to sleep.