Is SVS seeing files that you cannot?

Q:
Roie asked, "In my teaching I run a lot of programs from the Windows command line. (Hmmm, wonder how I will do that in Vista?) I have set up a data layer with folders for all my class labs, and an app layer for each development tool -- Dev-C++ for example.

This morning, I was working up a demo C++ program that creates a local file. Dev-C++ and the lab data area are both activated. Dev-C++ is building program files correctly and I can see them fine. The demo program is run from the command prompt and seems to run fine. It writes a named file, closes it, reopens it and reads what was written and prints it back out. Everything works fine. Only problem is that the new file does not show up in the directory listing - even though the application can find it. In searching through the layer properties, I see the new file in the folder as part of the RW layer, and the program files created by Dev-C++ are all in the RO layer. The application can see the new file, but I cannot from the command line or from Windows Explorer.

Any ideas why?

A:
Hi Roie, you've found one of our favorite bugs in SVS 2.0 -- I thought I was the only one who had found this!

Here's the issue: In SVS 2.0, executables running from a data layer make file system and registry changes to the writeable sublayer of that data layer, just as if it were an application layer. Problem: SVS does not display that writeable sublayer to the system, only to the process tree that's running out of the data layer.

The "data" you are outputting into the data layer is executable code. Then, when the output executable runs, you get the behavior you describe.

This has been fixed in SVS 2.1. I suggest getting Beta 2 and running this scenario again.