EclipseIO::writeTimestep ensure summary data.

[joakim-hove]

There was an incorrect early return in the writeTimestep( ) routine. The consequence of that was that summary data was not written for sub-report time resolution. This will require reference data updates:

1. The evaluation of total variables WOPT = \int_{t_1}^{t_2} dt WOPR which will be integrated over a complete report step, instead of subdividing the integeral over the timesteps of the simulator.
2. There will be too few time steps in the summary data file.

[joakim-hove]

jenkins build this with downstreams update_data please

[joakim-hove]

jenkins run norne spe polymer please update_data please

[joakim-hove]

jenkins build this with downstreams update_data please

[joakim-hove]

@akva2: The PR: https://github.com/OPM/opm-data/issues/178 looks fine, but the changes in this PR will mandate updates also to to the norne++ reference data - is that within scope of the update_data attribute?

[joakim-hove]

jenkins build this with downstreams opm-data=178 please

[joakim-hove]

jenkins build this with downstreams opm-data=178 please

[joakim-hove]

jenkins build this with downstreams opm-data=177 please

[akva2]

currently not supported, not necessarily out of scope but i have only considered regression tests for now.

[joakim-hove]

Ok - I'll go the manual route then.

[joakim-hove]

Ok - I'll go the manual route then.

Ehhh - is it possible to apply for a opm-data PR with updates to all the reference data[1]? As someone not doing this on a regular basis (and I guess I am actually number two on the list - I must admit I find this quite heavy going),

[1]: The summary file CASE.UNSMRY should be updated - might be unmodified in some of the cases.

[akva2]

which test doesn't update the unsmr file ?

ref https://github.com/OPM/opm-simulators/blob/master/tests/update_reference_data.sh#L23
only the norne init test skips it afaict.

[joakim-hove]

which test doesn't update the unsmr file ?

All tests update it - but for this particular PR it might be unchanged for a couple of the tests.

[akva2]

then i simply do not understand what you are asking for.

[joakim-hove]

then i simply do not understand what you are asking for.

Just assistance with the manual grunt work / book-keeping required to run all the integration tests and update the reference data, without f...ing up somewhere.

[akva2]

status for the integration tests:

• polymer compares to eclipse data.
• spe compare to eclipse data, as well as opm data which is updated through the regression tests.
• norne plots a reference solution for opm along with eclipse and the executed runs. i have added the necessary fiddlery to support updating the reference file through the trigger in added: update norne integration test from jenkins trigger opm-simulators#1044 but that will need @alfbr approval before i merge. an update should not be strictly required due to this PR as i guess it will plot the same data as before.

[joakim-hove]

status for the integration tests:

OK - thank you for the update.

I'll merge - and take the flak if there is one.

[joakim-hove]

jenkins build this with downstreams update_data please

[joakim-hove]

jenkins build this with downstreams opm-data=179 please