If steps 1-8 outlined in my previous post actually work correctly 100% of the time for anyone on Fremantle, could you please speak up with your firmware version? (Failure reports would be appreciated as well, but success reports would be more puzzling.) Bingo. What we're supposed to see after "MicroB ready" is this: Code: Checking to see if we should kill MicroB Checking to see if we should kill MicroB Closing MicroB (which has been observed at least once in the wild -- which makes this report rather troubling). Is any documented signal sent when the last window closes? Not that I know of. I'm relying on behavior gleaned from tracing MicroB in the SDK to determine when the last browser window closes. Clearly, at least in your case, that behavior is not happening. Can you repeat steps 1 to 6, and this time, find the PIDs for browserd ("ps -ef | grep browserd") before step 4
Checking to see if we should kill MicroB Checking to see if we should kill MicroB Closing MicroB
~ $ ps -ef |grep browserd 1205 user 40232 S /usr/sbin/browserd -d 1616 user 60128 S /usr/sbin/browserd -s 1616 -n RTComMessagingServer 1692 user 2092 S grep browserd
, after step 4,
~ $ ps -ef |grep browserd 1205 user 40232 S /usr/sbin/browserd -d 1616 user 60128 S /usr/sbin/browserd -s 1616 -n RTComMessagingServer 1706 user 81260 S /usr/sbin/browserd -s 1706 -n browserui 1723 user 2092 S grep browserd
and after step 6?
~ $ ps -ef |grep browserd 1205 user 40232 S /usr/sbin/browserd -d 1616 user 60128 S /usr/sbin/browserd -s 1616 -n RTComMessagingServer 1737 user 55720 S /usr/sbin/browserd -s 1737 -n browserui 1743 user 2092 S grep browserd
(In the SDK, there are two browserds after step 4, and one of them closes and is replaced after step 6; it's this change that I have several ways of looking for. If this isn't happening consistently on the device, we're back to square 1 ...)