Visual Studio 2010 SP1 debuuging problems

[David Stringer]

After a number of observations from my associate developers and more careful testing by myself I am convinced we have serious problems regarding the use of Visual Studio 2010 SP1 and VirtualBox when debugging.

The typical guest operating system is Windows 7 x64 with VT-x/AMD-V obviously enabled.

When attempting to debug Visual Studio 2010 solutions we get one of the following two problems:

1. An exception "Attempted to read or write protected memory. This is often an indication that other memory is corrupt" occurs regularly indicating some sort of attempted deep memory corruption in the process space.

2. The debugger randomly skips onto future break points when issueing step into requests. You can reliably produce this effect if you hit the step into key (F11) rapidly. This makes complex debugging almost impossible.

After much checking with the settings of the virtual machines I may have been able to narrow it down to one setting that may improve the reliability of the virtual machine.

That setting is change the virtual machine to use a single processor only, in which case at first investigation the problems appear to go away. This may indicate some kind of threading problem between Visual Studio 2010 SP1 and Virtual Box.

For your reference we observed this same problem with Visual Studio 2012 RC, but thought that it had gone away with Visual Studio 2012 RTM but cannot be sure as we probably didn't debug that much following the RTM upgrade.

I have converted one test machine to VMWare format and tested under VMWare Workstation 9.0.2 and shown that problem is not exhibited under VMWare Workstation with multiple processors enabled.

Therefore I conclude it is some sort of threading issue between Visual Studio 2010 SP1 and VirtualBox.

Hope this can fixed ASAP as I'm probably going have to move to VMWare workstation otherwise.

[pgates]

I have this problem too. The most active ticket for this issue appears to be https://www.virtualbox.org/ticket/9659.

[Frank Mehnert]

Yes, I think it's the same problem like #9659. Please try the test build I've provided there.