Open dezidizon opened 2 months ago
@hxiedwr @dwr-ycheng @dwr-zroy -- I'm including you in this discussion because you were on the other testing thread.
In a recent status call, I mentioned that Josh and I have been informally doing "smoke tests" to catch obvious problems with installer builds before sending them to DWR for more extensive tests. Over time the smoke tests have become more consistent and we've been following checklists like this one:
For the run testing step, I've been using a CalLite model that was provided to us in March, labeled "CalLite 4.1 TF." For the smoke tests I just run the model for one year so that it completes in under a minute.
Additionally, if we're working on a particular problem we check that the fixes that have been checked in make the changes that we expect.
Do you see any gaps in this? Is that CalLite model an appropriate choice for this purpose? It runs the CBC solver, so we're not screening for problems any others.
We usually don't change workspace but load the study into WRIMS GUI.
I think using CalLite model to test on your side is good. It is just used to prove the concept. Once the new package is forwarded to us, we will test the package with cs3, which takes much longer time.
I prefer RMA not to spend a lot of time on testing.
Smoke test of the WRIMS v2.2 installer