After some troubleshooting we've found that it is a socket issue. Normally we connect to our Daemon via a socket. When this is done from command line it works fine. When the socket connection is initiated from the JNI calls the connection request times out. Is there a subtlety we've overlooked about sockets and permissions?
On Nov 16, 1:58 pm, Abhinayak Mishra <[email protected]> wrote: > from command line are you passing in any parameters? what about the call > from JNI? Is it possible the calls from the JNI don't get properly sent to > the daemon and are malformed somehow? Simple test if you can recompile your > daemon would be to print out what it gets in the two instances. > > -Abhinayak > > On Mon, Nov 16, 2009 at 1:11 PM, Abram <[email protected]> wrote: > > So I have a daemon running and I can talk to it/use it just fine from > > command line. When I try and make calls from JNI to the daemon it's > > like I'm talking to the code, but not the code the daemon is talking > > to. That's kind of confusing.... > > > The daemon works with an SDIO card and is how we route calls to the > > device. Using command line, I ask the daemon for the device and all is > > well, the daemon can find it and return information. When making the > > calls from JNI the daemon says the device is not there. I can see all > > the calls made in logcat. But there is a night and day difference > > between JNI and command line. > > > Currently the daemon is being started by hand in the Alt-F1 console > > window after android has loaded. > > > Is there something that I am missing between calls made from JNI to a > > daemon service and calls made from command line to a daemon service? > > It almost seems like I'm talking to two different instances of the > > code base. > > > -- > > unsubscribe: > > [email protected]<android-porting%[email protected]> > > website:http://groups.google.com/group/android-porting -- unsubscribe: [email protected] website: http://groups.google.com/group/android-porting
