I’ll steal Steve’s
format…..
Component Model Interoperability
testing
Goal:
To achieve deployment-time interoperability
of the component model frameworks written by different people.
Means of achievement:
Testing
Means of testing:
Pre-defined components and
CDL templates, with measurable outputs.
Process:
A core set of test
components are created:
1)
A simple component with a predefined
resource property <testvalue> defined in its CDL. Additionally,
this component should expose resource properties for timing conditions, and its
state transitions. These would include <timeOfInitialize>,
<timeOfFail>, …, <stateList>. The <stateList>
would be a comma separated list of the transitions the component took in its
lifetime: initialized, running, failed, terminated, destroyed.
2)
A component which exhibits
failure conditions.
a.
Fails on transition to a
specified state.
b.
Fails on change of a
resource property.
c.
Faults on change of a
resource property.
3)
A component which can test
the value(s) of other component’s properties and assert truth. This
component is ideally a delegate, thus the remainder of the system which it
tests would act as a separate child system. This component should also
receive events and be able to assert conditions based on the event.
Each implementer creates a
test harness that uses their own Deployment API client to test these components.
No actual deployment is expected to be done, since these are not tests of
deployment, but of the infrastructure. The tests to be implemented follow
the component model test plan.
In order to do this, I will
need to update the Component Model Test Plan to fully specify these components
and the explicit positive and negative test cases. If this methodology is
reasonable, I will start that process.
Stuart
Stuart Schaefer
CTO
sschaefer@softricity.com
Softricity, Inc.
ph 617.695.0336 x112
fx 617.338.7769
www.softricity.com –
Virtually, Anything Is Possible™
The Softricity Desktop™ enables applications to be as
instantly available and easy to use as electricity
"All problems in computer science can be solved by another level
of indirection". -- Butler Lampson