FWD » Core Development » Runtime Infrastructure

The video is not enough to make a guess what happens here. Please add directory.xml, test program to reproduce and logs.

Șerban, please note that your directory.xml has a remote debug option

          <node class="string" name="jvmArgs">

            <node-attribute name="value" value="-Xmx512m -Djava.awt.headless=true -Xdebug -Xnoagent -Djava.compiler=NONE -Xrunjdwp:transport=dt_socket,address=2998,server=y,suspend=n"/>

          </node>

25/08/08 15:07:39.591+0300 | WARNING | com.goldencode.p2j.util.Utils | ThreadName:qtp1636257629-87 | not available host:port localhost/127.0.0.1:7449.
25/08/08 15:07:39.591+0300 | WARNING | com.goldencode.p2j.main.WebClientBuilderParameters | ThreadName:qtp1636257629-87 | DEBUG MODE: delayBetweenTriesToConnect=5000
25/08/08 15:07:39.591+0300 | WARNING | com.goldencode.p2j.main.WebClientBuilderParameters | ThreadName:qtp1636257629-87 | DEBUG MODE: delayBetweenPingTries=1000
25/08/08 15:07:39.591+0300 | WARNING | com.goldencode.p2j.main.WebClientBuilderParameters | ThreadName:qtp1636257629-87 | DEBUG MODE: embWebServerMaxIdleTimeout=0
25/08/08 15:07:39.591+0300 | WARNING | com.goldencode.p2j.main.WebClientBuilderParameters | ThreadName:qtp1636257629-87 | DEBUG MODE: enableDebugLogging=true
25/08/08 15:07:39.591+0300 | WARNING | com.goldencode.p2j.main.WebClientBuilderParameters | ThreadName:qtp1636257629-87 | DEBUG MODE: maxIdleTime=1200000
25/08/08 15:07:39.591+0300 | WARNING | com.goldencode.p2j.main.WebClientBuilderParameters | ThreadName:qtp1636257629-87 | DEBUG MODE: maxLostPings=6
25/08/08 15:07:39.591+0300 | WARNING | com.goldencode.p2j.main.WebClientBuilderParameters | ThreadName:qtp1636257629-87 | DEBUG MODE: webSocketTimeout=1200000
25/08/08 15:07:39.591+0300 | WARNING | com.goldencode.p2j.main.WebClientBuilderParameters | ThreadName:qtp1636257629-87 | DEBUG MODE: pingPongInterval=1200000
25/08/08 15:07:39.591+0300 | WARNING | com.goldencode.p2j.main.WebClientBuilderParameters | ThreadName:qtp1636257629-87 | DEBUG MODE: watchdogTimeout=2400000

But in your log there is this severe message

25/08/08 15:07:52.076+0300 |  SEVERE | com.goldencode.p2j.util.TransactionManager [TransactionManager.deferError] | ThreadName:Conversation [00000002:bogus], Session:00000002, ThreadId:00000005, User:bogus | <depth = 1; trans_level = -1; trans_label = null; rollback_scope = -1; rollback_label = null; rollback_pending = false; in_quit = false; retry_scope = -1; retry_label = null; ignore_err = false> [label = startup; type = UNKNOWN; full = false; trans_level = NO_TRANSACTION; external = true; top_level = true; loop = false; loop_protection = true; had_pause = true; endkey_retry = false; next_or_leave = leave; is_retry = false; needs_retry = false; FOR (aggressive) flushing = false; ilp_count = -1; pending_break = false; database_trigger = false; properties = ''; finally = none] Error processing finished; deferring throw until block/iteration completes
com.goldencode.p2j.util.EndConditionException: END-ERROR key function
    at com.goldencode.p2j.ui.chui.ThinClient.processCondition(ThinClient.java:23781)
    at com.goldencode.p2j.ui.chui.ThinClient.pause(ThinClient.java:10784)
    at com.goldencode.p2j.ui.chui.ThinClient.pauseBeforeEnd(ThinClient.java:19817)
    at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
    at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:77)
    at java.base/jdk.internal.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
    at java.base/java.lang.reflect.Method.invoke(Method.java:569)
    at com.goldencode.p2j.util.MethodInvoker.invoke(MethodInvoker.java:126)
    at com.goldencode.p2j.net.Dispatcher.processInbound(Dispatcher.java:808)
    at com.goldencode.p2j.net.Conversation.block(Conversation.java:418)
    at com.goldencode.p2j.net.Conversation.waitMessage(Conversation.java:344)
    at com.goldencode.p2j.net.Queue.transactImpl(Queue.java:1239)
    at com.goldencode.p2j.net.Queue.transact(Queue.java:687)
    at com.goldencode.p2j.net.BaseSession.transact(BaseSession.java:276)
    at com.goldencode.p2j.net.HighLevelObject.transact(HighLevelObject.java:221)
    at com.goldencode.p2j.net.RemoteObject$RemoteAccess.invokeCore(RemoteObject.java:1468)
    at com.goldencode.p2j.net.InvocationStub.invoke(InvocationStub.java:144)
    at jdk.proxy2/jdk.proxy2.$Proxy11.standardEntry(Unknown Source)
    at com.goldencode.p2j.main.ClientCore.start(ClientCore.java:476)
    at com.goldencode.p2j.main.ClientDriver.start(ClientDriver.java:287)
    at com.goldencode.p2j.main.CommonDriver.process(CommonDriver.java:522)
    at com.goldencode.p2j.main.ClientDriver.process(ClientDriver.java:171)
    at com.goldencode.p2j.main.ClientDriver.main(ClientDriver.java:330)
25/08/08 15:07:52.077+0300 |  SEVERE | com.goldencode.p2j.util.TransactionManager [TransactionManager.handleDeferredError] | ThreadName:Conversation [00000002:bogus], Session:00000002, ThreadId:00000005, User:bogus | <depth = 0; trans_level = -1; trans_label = null; rollback_scope = -1; rollback_label = null; rollback_pending = false; in_quit = false; retry_scope = -1; retry_label = null; ignore_err = false>  Throwing deferred error (END-ERROR key function)
25/08/08 15:07:52.455+0300 | WARNING | com.goldencode.p2j.security.SecurityContext | ThreadName:Reader [00000002:bogus], Session:00000002, ThreadId:00000006, User:bogus | SecurityContext.cleanupWorker did not complete properly - the following tokens are still in use: [com.goldencode.p2j.util.Utils$ContextContainer, com.goldencode.p2j.util.SessionUtils$1, com.goldencode.p2j.util.SessionUtils$2, com.goldencode.p2j.util.osresource.OSResourceManager$1]

Sergey Ivanovskiy wrote:

But in your log there is this severe message
[...]

That's an exception generated from the ESC key to end the application. What should happen is that on ESC press the app should terminate, but I need to wait for the websocket connection popup to appear and then press ESC again to close it.

On Swing ESC + Spacebar terminates and closes a FWD client instantly.

Another weird discovery is that ESC + Spacebar works perfectly fine on Incognito for me. It's instant and I don't have to wait for any popup.

Șerban Bursuc wrote:

Sergey Ivanovskiy wrote:

But in your log there is this severe message
[...]

That's an exception generated from the ESC key to end the application. What should happen is that on ESC press the app should terminate, but I need to wait for the websocket connection popup to appear and then press ESC again to close it.

On Swing ESC + Spacebar terminates and closes a FWD client instantly.

Another weird discovery is that ESC + Spacebar works perfectly fine on Incognito for me. It's instant and I don't have to wait for any popup.

I think that there is the bug here but it is not clear why this key action affects web clients.

but I need to wait for the websocket connection popup to appear and then press ESC again to close it.

I guess it is not correct. The application should terminates the java web client the same as for the swing client.

Pressing ESC twice should not cause the websocket connection lost message, same as holding ESC. This bug can be fixed to prevent waiting for the delayBetweenTriesToConnect timeout. It would be a good quality of life bug fix.

Șerban Bursuc wrote:

Pressing ESC twice should not cause the websocket connection lost message, same as holding ESC. This bug can be fixed to prevent waiting for the delayBetweenTriesToConnect timeout. It would be a good quality of life bug fix.

Please explain how it relates to the issue. The web client should act as the swing client and 4GL native application. delayBetweenTriesToConnect timeout should not be a solution for this issue.

Sergey Ivanovskiy wrote:

Șerban Bursuc wrote:

Pressing ESC twice should not cause the websocket connection lost message, same as holding ESC. This bug can be fixed to prevent waiting for the delayBetweenTriesToConnect timeout. It would be a good quality of life bug fix.

Please explain how it relates to the issue. The web client should act as the swing client and 4GL native application. delayBetweenTriesToConnect timeout should not be a solution for this issue.

What I mean here is on a web client application ESC -> ESC (press twice) or just holding ESC should immediately quit the application. It doesn't like in OE. It hangs and you need to wait delayBetweenTriesToConnect time for the "websocket connection lost" popup to appear. Only then you can press another ESC to safely free the port.

• The server hangs from a FWD bug or due to poor performance and you want to quit out.
• If you refresh the page and try to login again, the client port is still in use.
• You kill ServerDriver with a command like this: ps aux | grep ServerDriver | awk '{print $2}' | head -n 2 | xargs kill -9
• Now you can login, but midway through the loading screen disappears and you get the "no ports available" error again.
• You have to restart the server to be able to login again.

You should be able to at least kill the client thread and be able to relogin, or a page refresh should release the port. This wastes a lot of time in large projects when you are profiling for example, you sample for a few minutes but the scenario literally takes hours, so you want to close the client and relogin but you need to do a server restart.

An easy example of a hanging application is #10087-19, this testcase on the web client hangs and there is no way to end it besides killing the server, logging in again, getting the ports error, and then killing the server again.

edit: not a good example

The ESC issue is fixed by this change:

=== modified file 'src/com/goldencode/p2j/ui/chui/ThinClient.java'
--- old/src/com/goldencode/p2j/ui/chui/ThinClient.java    2025-07-28 08:11:21 +0000
+++ new/src/com/goldencode/p2j/ui/chui/ThinClient.java    2025-08-13 10:47:59 +0000
@@ -23860,7 +23860,7 @@
                                            lastKey);

       // END-ERROR action
-      if (action == Keyboard.KA_END_ERROR)
+      if (action == Keyboard.KA_END_ERROR && !inPause)
       {
          int effective = interactions;

Throwing an exception here results in the hanging state, but if no exception is thrown on the second ESC then the application ends gracefully.

It seems that the similar issue was discussed in some task but I did not find where it was.

I don't understand this try block:

11057.1.2   ca@gold |    
                    |          try
                    |          {
                    |             // honor the STOP, ENDKEY and END-ERROR actions
                    |             if (key != null)
                    |                processCondition(key.actionCode(), -1, true);
                    |          }
                    |          finally
                    |          {
                    |             // erase status message, only in ChUI
                    |             if (!noMessage && isChui())
                    |             {
                    |                wnd.window().status(null);
                    |             }
                    |    
                    |             if (isChui())
                    |             {
                    |                // hide messages
                    |                wnd.window().setMessageText(null, null);
                    |             }
                    |    

Why process the condition? Any key besides some modifier keys like CTRL and SHIFT end the application. It looks like we are deliberately trying to throw a StopConditionException here. I see #3100 with this change where there were some fixes to batch mode and app servers. Should this check not be done for GUI?