Puzzling access violation: SECURITY CHECK COOKIE ?

[WSinc]

This graphics program I am running crashes for no apparent reason,

at least according to the debugging printout I generate.

One puzzling thing, is that it does not always crash in the same location,

even when I don't change anything, inputs or anything else.

I do get a HEX code where it crashes, but how do I tell what routine is that in?

No source code is given, nor is there any traceback.

sometimes it crashes where it's calling "security check cookie."

Why would that routine be part of your software? Is this something the FBI

is pulling on us? I know no reason they would be spying on me,

but I have heard horror stories that they don't bother to get a court order any more.

I have the anti-virus stuff turned off, so I don't see why that would apply here.

[WSinc]

Here is another crash place. I upoladed another screen shot, a JPEG file. What is "Find PE scetion" Is that one of your routines? Like I said before, it does NOT always crash in the same place, even when NOTHING is changed about the way I run it.

[IanH]

For your first post - has the debugger dropped you into the C runtime library initialisation/termination code? There's a few lines of code there about setting up for exception handling that involves a security cookie. What's the name of the file that's shown in the debugger? You can sometimes end up in that C runtime code when something goes astray in you program that appears to confuse the default error handling that helps the VS debugger pinpoint the problematic line. I've not worked out what has to go wrong for you to end up there, but it is not a happy place. Your screen shot appears to show code from the rather low level (just above the kernel) of the base operating system (ntdll.dll). That is similarly not a happy place.

[WSinc]

Hi Ian; Well, is there a way (aside from smoking something from the garden) to make this a happier place? I used a mouse click to slow it down, and it doesn't crash then, and putting a delay loop in seems to help. But so far, I have not been able to narrow down the source of the problem. As I said earlier, the printout I put into a DEBUG file does not show anything suspicious. How would I get a storage map that I can use to determine which routine it happens in? assuming that is, that it is not misleading me entirely. The program runs hundreds of plot segments before it crashes. I am wondering if I upload the EXE if it would run in your environment. Or is there someone who can run it remotely? It requires an input data file (unformatted).