dive plan test

Dirk Hohndel dirk at hohndel.org
Sun Jul 5 11:47:06 PDT 2015


On Sun, Jul 05, 2015 at 08:20:21PM +0200, Robert C. Helling wrote:
> Dirk,
> 
> > On 05 Jul 2015, at 18:45, Dirk Hohndel <dirk at hohndel.org> wrote:
> > 
> > So I just pushed a commit that adds a TestPlan test.
> 
> right now, I cannot pull it. Are you sure you pushed it?

Um. Almost. Pushed now.

> I will sit down with the debugger tomorrow and try to find the origin of
> the differences. Consider this as taken care of.

Thank you. Much appreciated. I'm reasonably certain that I'm just missing
some setting that gets set up in the full app but does not get set up in
TestPlanner.

> I did not have time to compare the plans of the different programs, I
> hope they mostly agree where they should.

I didn't do much more testing there, either. Once I had a plan where there
was a discrepancy it seemed reasonably obvious that there might be
something wrong with my test application - no point in testing the plans
with something that clearly is somewhat broken.

> For a real test of the planner, we should not test with GF’s 100/100 as
> this is a very special case, better use 80/30 or anything else.
> Furthermore, we should also test plans in CCR and pSCR mode and with
> oxygen breaks enabled. But I guess, once we have the infrastructure for
> one plan, adding the others will not be too hard.

Exactly. I wanted to get one test (well, set of tests - imperial and
metric) correct and then expand to test many more variations. Once the
data is collected it's quite simple to reliably and reproduceably test
dozens of plans.

/D



More information about the subsurface mailing list