怎样捕捉OxC0000005类型的内存访问冲突引起的异常

[laoluo9932]

我的一个程序调用了一个第三方的库，该库我是无法修改和调试的。在debug下运行，在调用该库的某个API时偶尔会抛异常，导致程序崩溃。提示框显示：OxC0000005访问0x********地址错误，大致就是这样的信息。
我想捕捉到这个异常，这样进程在运行时就不会崩溃了。可是我用C++的try catch(...)捕获不到，用__try __except也捕捉不到。
请各位大侠给指点一下，有什么办法能捕捉到这种内存访问错误引起的异常。

[jimette]

刚刚调用CoInitialize 引起了上述错误 更改了一下位置 错误没有 关注！

[zcchm]

http://stackoverflow.com/questions/993551/seh-access-violation-and-stack-guard-page To catch access violation exceptions, you need to enable SEH exception handling (C/C++ -> Code Generation -> Enable C++ Exceptions -> Yes, with SEH Exceptions). Be warned though, this will more likely than not make your software less reliable. You won't always get an access violation on every bad memory access -- it is very possible that your bad pointer points to valid virtual memory. In that case could overwrite any valid data in your program, but not see the consequences for a long time. You are much better off to ensure that you would never write to bad memory in the first place.

[漫步者、]

用assert试一试，或者debug-》 call back

[一个傻冒]

是不是栈溢出，覆盖了异常处理在栈中的记录，所以捕捉不到了？

[okgogo2000]

启动时 if (!ProcessShellCommand(cmdInfo)) 有时会crash报以上错误。 关注中。。。

[stjay]

这个能够捕捉0xC0000005异常

int SehFilter(DWORD dwExceptionCode)
{
	switch (dwExceptionCode)
	{
	case EXCEPTION_ACCESS_VIOLATION:
		return EXCEPTION_EXECUTE_HANDLER;
	}
	return EXCEPTION_CONTINUE_SEARCH;
}

int test()
{
	*(int*)0 = 0;
	return 0;
}

int main(int argc, char* argv[])
{
	__try
	{
		test();
		printf("test()");
	}
	__except (SehFilter(GetExceptionCode()))
	{
		printf("EXCEPTION_ACCESS_VIOLATION");
	}

	getch();
}