Inconsistent handling of PORV/MULTPV in EDIT section

[blattms]

We have a test PORV in FieldPropsTests.cpp with:

EDIT

BOX
  1 10 1 10 4 4 /

MULTPV
  100*4 /

ENDBOX

BOX
  1 10 1 10 3 3 /

PORV
  100*3 /

ENDBOX

which has the expected result, because when getting PORV from the FieldProperties for writing the first time, we will calculate it from PORO, NTG, MULTPV and cell volume.

But if we put PORV before MULTPV, then the multiplier will not be accounted for and the test breaks, see #4023. Due to the non-overlapping boxes this should definitely be a problem.

What if the boxes overlap or are the same? Would order matter here?

[yasser508]

That's a great question! Given how things behave with separate boxes (as mentioned regarding #4023), I'd definitely expect the order to matter for overlapping or identical boxes too.

My thinking is that the code probably just processes those EDIT commands one after the other as it reads them. So, if two commands affect the same cell because the boxes overlap:

• If you put MULTPV first, it'll multiply the pore volume. Then, when PORV comes along, it'll act on that already-multiplied value.
• If PORV is first, it sets the pore volume (and as we saw, it might miss the multiplier that comes later). Then MULTPV would multiply whatever value PORV just set.

Those two scenarios sound like they'd give different final pore volume results for the cells caught in the overlap. So yes, the order seems pretty crucial here as well.

It might be helpful to create a small test case with overlapping boxes just to confirm this is how it works.