Intermittant GPF

[judyh]

Hi all,

I am getting a really strange intermittant GPF from a program that we've had for a long time and that I am modifying.  Note that I'm using CVI 2009 for the first time to modify and build this code - I wonder if that may cause some problems with older UIR controls, but I have converted the UIR file to lab style controls.

What's really strange about this is that, typically if my code causes a GPF, it's pretty straight forward to debug and the issue it that it's trying to write to unallocated memory.  This one is saying it cannot read from a memory location.  It happens when the code populates and displays a report panel and it does not happen in the same place of the code each time.  Sometimes it occurs before the panel is displayed, sometimes afterward.  I have not been able to get this to happen in the debug build.  It only occurs in the release build.  Also, we've tried it on different machines just in case there was an issue with out test machine.  We did a clean install - removed all CVI programs and DLLs before installing on the test machine.  It's a really tough little bug to hit.  Any suggestions?

Thanks,

Judy

[nickb]

Hey Judy -

At first glance, it looks like that crash may be coming from inside the CVI Run-Time engine.  Do you see the standard windows crash dialog when this occurs?  If so, can you please follow the steps outlined on this page to generate a crash dump I can take a look at?  If you are able to generate a crash dump, please remember to include the map file for your application with the .dmp file (there are instructions for this at the bottom of the page).

Let me know if you can get these files, and I'll take a look as soon as possible.

NickB

National Instruments

[judyh]

Hi Nick,

I was unable to get the crash dump because we never get the Windows crash dialogue box.  However, I did uninstall CVI 2009 and reinstall CVI 9.0 and this appeared to get rid of the GPF.  Sounds like an issue with 2009's run time engine.  Another possibility is that the code is using outdated functions or controls - it's very old code and I've updated about half of it for this release.  Can you possibly point me toward a list of depricated or outdated functions/controls and I can check for them in this code?  That might help.  I know CVI 2010 is due out soon and when we receive that update, I'll check to see if this behavior is there.

Thanks,

Judy

[nickb]

Hey Judy -

Because we strive to maintain backwards compatibility with each new version of the CVI Run-Time Engine (apparently failing on occasion... ), we don't really have a list of outdated controls or functions that we know to cause problems like what you're seeing. 

Instead, I'd like to approach this a little differently.  Can you let me know the exact version of your run-time engine (9.1.x.xxx)?  Is there any way you could post code or an executable I would be able to reproduce this with?  What happens when you click cancel on that dialog - do you have a registered just-in-time debugger?

NickB

National Instruments