On Exit database method
On Exit database method
| Does not require any parameters |
|---|
This command is not thread-safe, it cannot be used in preemptive code.
The On Exit database method is called once when you quit a database.
This method is used in the following 4D environments:
- 4D in local mode
- 4D in remote mode
- 4D application compiled and merged with 4D VolumeDesktop
Note: The On Exit database method is NOT invoked by 4D Server. You should use On Server Shutdown database method in this context.
The On Exit database method is automatically invoked by 4D; unlike project methods, you cannot call this database method yourself by programming. You can however execute it from the Method editor. You can also use subroutines.
A database can be exited if any of the following occur:
- The user selects the menu command Quit from the Design Environment File menu or from the Application environment (Quit standard action)
- A call to the QUIT 4D command is issued
- A 4D Plug-in issues a call to the QUIT 4D entry point
No matter how the exit from the database was initiated, 4D performs the following actions:
- If there is no On Exit database method, 4D aborts each running process one by one, without distinction. If the user is performing data entry, the records will be cancelled and not saved.
- If there is an On Exit database method, 4D starts executing this method within a newly created local process. Note that 4D will eventually quit—the On Exit database method can perform all the cleanup or closing operations you want, but it cannot refuse the quit, and will at some point end.
When 4D is about to quit, the program:
- asks running processes to abort their execution as soon as possible (taken into account only in processes executing 4D code)
- sets their ‘process aborted’ flag
- awakes processes that have been delayed by DELAY PROCESS or other means.
4D continues iterating processes while any process is still alive, for a maximum duration of 13 seconds. To respond to a quit action, we recommend to set a flag (such as an entry in the Storage object) in the On Exit database method or in your custom quit method. It could make sense to loop in that method (max 10 seconds) to give time to your running processes to respond and quit themselves properly. It is not recommended to wait longer than 10 seconds, to avoid that 4D kills the process or the operating system kills the whole application.
The On Exit database method is the perfect place to:
- Save (locally, on disk) Preferences or Settings to be reused at the beginning of the next session in the
- Perform any other actions that you want to be done automatically each time a database is exited
Note: Don’t forget that the On Exit database method is a local/client process, so it cannot access the data file. Thus, if the On Exit database method performs a query or a sort, a 4D Client that is about to quit will "freeze" and actually will not quit. If you need to access data when a client quits the application, create a new global process from within the On Exit database method, which will be able to access the data file. In this case, be sure that the new process will terminate correctly before the end of the On Exit database method execution (by using Storage or a shared object, for example).
Note: In a client/server environment, the On Exit database method behaves differently depending on whether the user quits manually (through the Quit menu command or a call to the QUIT 4D command) or 4D Server is shutdown, which forces all clients to quit.
When 4D Server shuts down and provides a cutoff time (e.g. 10 minutes), each connected client displays a warning message and if the user quits during the given time frame, the On Exit database method is executed normally. However, in other cases (e.g. the user does not respond in time, the server requests to quit immediately, or the client is manually disconnected by the Administrator), the On Exit database method is executed at the same time that the connection to the server is closed. As a result, the code in the On Exit database method can't launch another local or server process, and can't wait for other processes to be cancelled (nor can these processes continue to access the server). If it tries to do so, a network error is generated (such as 10001 or 10002) since the connection to the server is already closed.
Note
The following example shows a typical case where you launch one or more background processes doing regular jobs, running endless, in the On Startup database method (or in the On Server Startup database method). A flag in the application's Storage object is used to tell them to finish – and to check if they have finished.
- In the On Startup database method
Use(Storage)
Storage.global:=New shared object("ExitNow";False)
End use
$p:=New process("Serverjob";0;"Background job#1")
$p1:=New process("Check_Invoices";0;"Background job#2")
- Here is an example of such a background method:
// do regular background jobs on the server
While(Not(Bool(Storage.global.ExitNow)))
// run accounting jobs, calculate data, etc
// ...
DELAY PROCESS(Current process;60*60) //sleep for one minute, then repeat loop
End while
- In the On Exit database method:
Use(Storage.global)
Storage.global.ExitNow:=True //ask all background workers to quit
End use
$time:=Current time
$finished:=False
While((($time+10)>(Current time+0))&(Not($finished)))  // max 10 seconds
$list:=Get process activity(Processes only) // get list of all processes
$sublist:=$list.processes.query("name='Background job@' and state >=0") //is there still a background job running?
$finished:=($sublist.length=0)
If($sublist.length>0)
For each($job;$sublist)
RESUME PROCESS($job.number) //make sure all processes are awake
End for each
End if
End while