Motorola C450 ユーザーズマニュアル

 

common error is to implement 

startApp

( ) to execute instructions that are only 

intended to be executed once during MIDlet execution. The correct implementation is to 
include in 

startApp

( ) those instructions which can and should be executed each time 

the MIDlet changes from the Paused state to the Active state. The same logic should be 
applied to 

pauseApp

( ). 

The code sample below shows implementation of using 

startApp

( ). 

startApp

( ) 

performs operations for the initial launch of the MIDlet as well as any operations that need 
to take place each time the MIDlet changes state from Paused to Active. Booleans are 

used to determine whether the MIDlet has started and whether it’s in the Active state. 
These Booleans can also be used by other MIDlet threads to determine state. 

package midp.demo; 

import javax.microedition.midlet.MIDlet; 

 

public class Demo extends MIDlet { 

// The MIDlet has been started already 

private boolean isStarted  = false;  

 

// The MIDlet is in active state  

public  boolean isActive   = false; 

// (in most cases these booleans are used by other threads) 

 

protected void destroyApp(boolean unconditional){ 

 

isActive  = false; 

} 

protected void pauseApp(){ 

 

isActive = false; 

} 

protected void startApp(){ 

 

isActive = true; 

 if 

(!isStarted){ 

 

 

//...Use getDisplay(), setCurrent(),  

 

 

// and other initial actions 

 

 

isStarted = true; 

 } 

} 

 

The MIDP specification of 

javax.microedition.midlet

 allows for some 

latitude in the implementation, therefore, it cannot be assumed that all MIDlets are 
perfectly compatible with all devices. Some MIDlets may execute flawlessly on desktop 

simulators such as Sun’s Wireless Toolkit and contain a gain/loose focus mechanism 
such that the MIDlet transitions between the Paused and Active states and 

startApp()

 and 

pauseApp()

 are called.