Hog3PC [] "class fps::exception" Error on New Show?

Just to add to this, when I press "details" at the error window I get:

Exception source callstack, desktop process, 3712 thread

My network details are:

Gigabit switch. Everything working fine as far as I can tell.

Hog3PC machine: 192.168.2.2 (default gateway 192.168.2.9).

Visualiser machine (although not being used): 192.168.2.1 (default gateway 192.168.2.9).

The good news is that your show is getting created properly. As you found, thiis error comes up on new show creation, but if you exit and re-launch the same show it's fine.

We've had a hard time reproducing this at HES. The next time you (or anyone) sees this, can you press the button for more info, and then write down *exactly* what it says? Or take a photo of the screen and post it. All those strange numbers will help us.

Thanks for the reply Eric (especially on a Saturday evening!)

That message is *exactly* what I am getting when I press "details":

Exception source callstack, desktop process, 3712 thread

It's good to hear that the show is being created properly though :) It seems to be something to do with the dp8 service (if that makes sense) - that is the point I get the error.

System is: Core2Duo Processor, 2GB RAM, 256MB shared video RAM. NVidia MCP73 chipset - the PC is an ACER Aspire M1640. It's a pile of s**t really, but I just wanted to make it useful by using it for Hog3PC!

I'm going to try it on another XP SP3 system tomorrow: Pentium D 930, 3GB RAM, ATI 256MB graphics, ASUS P5WD2 Premium MoBo (Intel Chipset).

Incidentally, I'm not sure if this has anything to do with it, but my advanced network TCP/IP settings are changed from the MS defaults. The changes I have made are:

- Enable LMHOSTS lookup is DISABLED
- NetBIOS settings are: Enable NetBIOS over TCP/IP

These settings are in place to make another program's license service run properly. The system I am trying tomorrow will have the default settings.

Cheers,

Martin

I'm glad I could answer your questions. There's usually a few of us who watch the forums on weekends, but it's not guaranteed. If you have a critical issue, calling support will usually get faster response.

Typically when you get that "class fps" error message, there should be a bunch of lines that look like this:
[0x0e4624f8] fps::exception::exception+0x365
[0x0fe468f2] PlayList:PlayList(unsigned ShowItem)
and so on for 15 or 20 lines. Do you see anything like that? Those are the message that will be the greatest help for us if you see the message again.

I don't think there's anything wrong with your computer. We've heard of this happening on a variety of consoles and Hog3PC systems. A similar error would occur very rarely on earlier software, but it seems to be happening more often on 3.0. We're looking into it.

[quote=ericthegeek]Do you see anything like that? Those are the message that will be the greatest help for us if you see the message again.

i noticed in the release notes a mention of stack trace dumps, but i've yet to see any. in the typical "class fps::exception" handler box, i do see a little colon after the thread id, suggesting that a stack trace should follow, but alas, i get a whole lot of nothing.

Thanks Eric. From memory I'm not getting anything except that one line, but when I try it on my other system later today, if I get anything I will take a screen dump.

Are these lines in a dump file somewhere?

M

You can look for any dump files in the LOG folder. See if there's one with the correct date.

All,

We are aware of this problem that occurs sometimes when starting a new show. It is logged as bug #12631 and is top of the list for an upcoming patch.

While annoying, the problem does not disrupt anything and as you found, starting another new show will resolve the problem.

For all the users experiencing the "class fps::exception" error for the desktop process:

We have confirmed and logged this bug (bug #12631) as a high priority issue and will be patching it in our next software release. This problem is known to be isolated to new show launches and is not 100% reproducable as it is the result of a race condition with the desktop process. Currently the workaround for this issue is just as Martin suggests:

Open windows task manager
Go to the processes tab
Exit ALL the processes (the ones ending in -golden.exe) in task manager
Relaunch Hog3PC
Reopen the new show that you just created.

Just to confirm this has not happened on the third system that I have tried. I don't know if it helps, but this is the only PC I am using with the default LAN settings (see post #4 for details).

Cheers,

Martin

Purely out of interest... I've seen it mentioned a couple of time by HES guys on here. What is a "race condition"?

M

Race Conditions are when different processes or threads try to update the same data structure at the same time. The result can be partly what one thread wrote and partly what the other thread wrote. This garbles the data structure, typically causing the next thread that tries to use it to crash. In this case it's the desktop.

[quote=c_muenchow]This problem is known to be isolated to new show launches and is not 100% reproducable

i think dp8k is trying to get in on this; when i change my desk's user number, not only does the common problem occur, but i sometimes get a nice VC runtime error too.

if i'm listening, it says to me:
[code]
From dp8k process, pid 4816, thread 4816
[11722389 dp8k 4816] Error: Trap: '!m_pMessageQueue->Post( smsg, SysID_Launcher )' is TRUE
File: .\dp8k_app.cpp (Dec 11 2008) , Line: 190
Version: v3.0.0 (b 2417)[/code]

the most recent minidump was taken hours before that, but it looks about right:
[code]
Unhandled exception at 0x77c13592 (ntdll.dll) in dp8k-win32-golden.exe.dmp: 0xC0000005: Access violation writing location 0x00000008.

> ntdll.dll!_RtlEnterCriticalSection@4() + 0x12 bytes
dp8k-win32-golden.exe!PegWindowProc() + 0xafd0d bytes
[Frames below may be incorrect and/or missing, no symbols loaded for dp8k-win32-golden.exe]
dp8k-win32-golden.exe!PegWindowProc() + 0xae00 bytes
dp8k-win32-golden.exe!PegWindowProc() + 0xb63f4 bytes
dp8k-win32-golden.exe!PegWindowProc() + 0xc2adc bytes
dp8k-win32-golden.exe!PegWindowProc() + 0xc2f89 bytes
[/code]

it'll happen at the same moment that the common problem occurs, and i've never seen it happen alone.

Thanks for the stack traces. A fix for this issue went into test yesterday. It should be available with the first patch for 3.0.0. I can't give a timeframe for the patch at this time though.

In the meantime, restarting the newly created show will get you up and running.