Thank you for your help, in my case is a Nexus One v2.3.6 (a bit old) and I have been doing a bit more of testing, this is the logcat when the access point change:
03-29 08:33:12.890 I/wpa_supplicant( 934): CTRL-EVENT-DISCONNECTED - Disconnect event - remove keys 03-29 08:33:12.890 V/WifiMonitor( 96): Event [CTRL-EVENT- DISCONNECTED - Disconnect event - remove keys] 03-29 08:33:12.890 V/WifiStateTracker( 96): New network state is DISCONNECTED 03-29 08:33:12.900 I/wpa_supplicant( 934): CTRL-EVENT-STATE-CHANGE id=1 state=0 BSSID=00:00:00:00:00:00 03-29 08:33:12.900 V/WifiMonitor( 96): Event [CTRL-EVENT-STATE- CHANGE id=1 state=0 BSSID=00:00:00:00:00:00] 03-29 08:33:12.900 V/WifiStateTracker( 96): Changing supplicant state: COMPLETED ==> DISCONNECTED 03-29 08:33:12.900 D/ARIADNA ( 606): connectionChangeReceiver - SUPPLICANT_STATE_CHANGED_ACTION 03-29 08:33:12.900 D/ARIADNA ( 606): Stop Socket and start buffering... The problem is that the last message received by the server is on 08:32:54 after that 10 messages are sended but not received (before the broadcast received). ¿Is there any way to detect if the message has arrive to the server? ¿Other way to test connectivity or test the server? this is the way I implement the socket: sock = new Socket(); sock.bind(null); InetSocketAddress isa = new InetSocketAddress(serverIPAddress,serverPort); sock.connect(isa, conectionTimeOut); sock.setKeepAlive(true); isR = new InputStreamReader(sock.getInputStream(), "UTF8"); reader = new BufferedReader(isR); out = new BufferedWriter(new OutputStreamWriter(sock.getOutputStream(), "UTF8")); Thank you so much On 29 mar, 00:52, Irfan Sheriff <isher...@gmail.com> wrote: > > It would be good to know if there are devices beyond ICS that still have > > issues around not resetting socket connections when a network goes away. > > > Beyond ICS? I don't know what this means. > > > My "flagship" Galaxy Nexus still hasn't received 4.0.3. > > > And what do you mean by "would be good to know"? > > > Should we, developers, and our users, moonlight as Android's test > > department? > > I think we dont have a CTS around this behavior. I will look at filing a > bug internally for this to ensure things remain sane going forward. > > Thanks > > > > > > > > > > > Ideally, apps should not have to listen to any broadcasts to handle > > network switches and should get resets on an existing socket. > > > Not ideally -- really -- apps are supposed to work on more than the 1.6% > > of devices with ICS. > > > And "beyond ICS" - what portion of the market is that, at this time, > > exactly? > > > -- K > > > Irfan > > > -- > > You received this message because you are subscribed to the Google > > Groups "Android Developers" group. > > To post to this group, send email to android-developers@googlegroups.com > > To unsubscribe from this group, send email to > > android-developers+unsubscr...@googlegroups.com > > For more options, visit this group at > >http://groups.google.com/group/android-developers?hl=en -- You received this message because you are subscribed to the Google Groups "Android Developers" group. To post to this group, send email to android-developers@googlegroups.com To unsubscribe from this group, send email to android-developers+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/android-developers?hl=en