I wonder whether the timer in the applet itself should be running if we fail to scan, I think it should be stopped when we fail and resumed again when a successful scan was done. If I do get it correctly and we fail to do a scan, we schedule a new one in e.g. 5 seconds, after 5 seconds we perform a successful one, but after another 5 seconds this timer will perform a new one and fail again. Stopping this timer when we fail to scan and resuming it later will result into higher ration of successful scans, am I right?

This timer never gets deleted.

timer will be one of the value of hash m_wirelessScanRetryTimer, which are cleaned line 516.
I could be missing a delete there though such as delete m_wirelessScanRetryTimer.remove(interface)

What about this?

You remove it just from the map, but it doesn't get deleted, you would need to do something like

delete m_wirelessScanRetryTime.take(interface);

I removed this because

1:
Request scan failure should never happen anymore because the handler makes sure never to call requestScan too often.
The timer can just constantly ask for request scans it does not impact the "real" requestScan sent.

2:
Now the scans are per interface, meaning we check the interface states by interface and they could have different state.
It does not match well with this timer : if an interface fails, should we stop to scan the others ? I believe not.
A solution would be to have timer by interface, but we already have this in the handler anyway so it seems overkill to do it here.

@jgrulich Any feedback ?
You approved this a while back but we still have this point to deal with.