1.3.11 Debugging Issue?

[terrymcg]

Hi Folks...

Since the 1.3.11 update, I'm having trouble debugging modules.

Specifically, after I've run a module once or twice from within VMD, it won't run any modules (including the samples) until I restart my system (Win10).

The build runs, the output screen is cleared, but the window for the test platform never opens and the Build menu stays in "Stop Running" mode.

I'm certain this is new to 1.3.11 (I didn't spend any time with 1.3.10, got both updates in series), and wonder if anyone else is having a similar issue.

Having to restart between runs certainly lengthens the edit-compile-debug cycle.

Cheers,
--
Terry McG

[CherryRuss]

Hi Terry.

Sorry you're having this problem. Is it happening with both the "Debug" and "Run without Debugger" options or with just one of them?

When you have the chance, please get the module designer into this state and then send us your log files. If you go to the "Miscellaneous" page of the preferences dialog and click the "Open Log Files Folder", it'll open the folder in Windows Explorer. The files will be called ModuleDesigner.log (for the module designer) and ModuleDesignerTestClient.log (for the debugger client). Try clicking the "Stop Running" menu option first to see if it'll kill the child app (it might take 5 - 10 seconds for the app to die, should it succeed) to help prevent the client log from getting truncated.

Thanks,
Russ

[terrymcg]

Hi Russ...

Thanks for the quick reply!

Yes, it happens both with and without the debugger. "Stop Running" *will* (eventually) bring VM back to life.

I've sent a link to a collection of log files via the support form.

The "Good" and "Bad" logs are from back-to-back runs right after a restart. The other two are pre-restart (things stalled last night and I hadn't yet restarted).

Thanks in advance for your help!

Cheers,
--
Terry McG

[terrymcg]

Hi Folks...

FWIW, I've "solved" the problem by rolling back to 1.3.5 (for both VMD and VM).

It was hit-n-miss for me with 1.3.11... sometimes things would be fine for a while, then the "debug shell fails to start" issue I mentioned above would return... eventually with a vengeance:

On Friday there was a Java update, and by Saturday I couldn't run anything in VMD, regardless of reboots. After messing around with Java versions w/o success, I eventually uninstalled Java, VMD and VM, then re-installed VM (1.3.5), VMD(1.3.5) and Java (8v211 SDK and RE).

This seems a stable config for me and I haven't seen the issue since switching to it.

Cheers,
--
Terry McG

[CherryRuss]

Hi Terry.

Thanks for the update. Your log files indicate that the java machine fails to initialize in the debug client, and we suspect it's due to the debugger from the previous instance failing to release its TCP socket, but we're not sure. We're probably going to have to add yet more logging to help us try to isolate it. We have edited the code so that if we're correct about the socket issue, at least the "Run without debugger" option should still work even when your debugger is borked. Sorry for the delays.

[terrymcg]

Hi Russ...

Thanks for the info - it sounds like you're on the right track, given what I saw in the logs.

If there's anything you need/want me to do to help isolate the problem, I'll be happy to give it a go.

Cheers,
--
Terry McG

[terrymcg]

Just to close the loop on this, the 1.3.13 update seems to have resolved the problem for me, so I'm back to current.

Thanks Cherry People!

Cheers,
--
Terry McG