[CMake] Random BAD_COMMAND error when running CTest

[Ian Mackenzie]

On Mon, Oct 20, 2008 at 10:53 AM, David Cole <david.cole at kitware.com> wrote:
> Another common source of apparently random results is uninitialized
> variables. I don't think it is in ctest itself, as this is the first we've
> heard of something like this on this mailing list (although anything's
> possible...)
> Have you run your test code through valgrind to see if it catches any
> uninitialized data access?
>
> David

Just tried with valgrind...all of the tests run cleanly.
Interestingly, when I run ctest itself through valgrind (e.g.,
'valgrind ctest' instead of 'valgrind bin/scalar_value_tests'), it
runs perfectly every time, and when I go back to running it on its
own, it starts failing again.

Any thoughts, or should I file a bug report with my distribution to
see if it's a library dependency problem?  It doesn't seem to be with
the distribution's packaging of CMake itself, since as I mentioned
before the binary version from the CMake website does the same thing -
although if I run the binary version through valgrind, I get a "still
reachable" memory leak message, while running my distribution's
version through valgrind I get no errors or warnings at all.

-Ian