J
jlea
How many times have we had a program that works toot-sweet in debug but dies
in release? Too many.
I've created .NET assembly containing a bunch of managed wrapper C++ classes
that wrap unmanged C++ classes; these unmanaged C++ classes are located in a
Win32 library file. The library is built using MFC as shared in debug but
static in release (this is how we typically to things).
The managed wrapper classes do a new on the unmanged class in the
constructor and then have pass through functions to the unmanaged class.
I have a console application that tests the wrapper assembly and all is
heavenly sweet in debug.
In release mode, the heap is corrupted before any lines are executed in the
constructor of any managed wrapper class; it is fine in the test application
right before it does a new on the managed wrapper class. _crtheap is zero in
malloc.c.
Anyone have any ideas?
Thanks.
Jon.
in release? Too many.
I've created .NET assembly containing a bunch of managed wrapper C++ classes
that wrap unmanged C++ classes; these unmanaged C++ classes are located in a
Win32 library file. The library is built using MFC as shared in debug but
static in release (this is how we typically to things).
The managed wrapper classes do a new on the unmanged class in the
constructor and then have pass through functions to the unmanaged class.
I have a console application that tests the wrapper assembly and all is
heavenly sweet in debug.
In release mode, the heap is corrupted before any lines are executed in the
constructor of any managed wrapper class; it is fine in the test application
right before it does a new on the managed wrapper class. _crtheap is zero in
malloc.c.
Anyone have any ideas?
Thanks.
Jon.