Sum is a sum

[aulemahal]

Pull Request Checklist:

• This PR addresses an already opened issue (for bug fixes / features)
  • This PR fixes Which units should a "sum" aggregation return ? #1645
• Tests for the changes have been added (for bug fixes / features)
  • (If applicable) Documentation has been added / updated (for bug fixes / features)
• CHANGES.rst has been updated (with summary of main changes)
  • Link to issue (:issue:number) and pull request (:pull:number) has been added

What kind of change does this PR introduce?

• to_agg_units will not change the units when the aggregation is "sum".
• Fixed the units when the initial data has a temporal dimensionality (ex "m/s") and an "integral" is performed. The output units were previously not reduced, yielding things like "d m s-1".

Does this PR introduce a breaking change?

Yes, but select_resample_op is rarely used by the user directly. And we could consider that this is in fact a fix from a previous ambiguous situation. No internal calls to to_agg_units were using "sum".

[aulemahal]

@Zeitsperre All tests fixed. This is ready to go. It will solve the remaining bug in the rolling statistics PR.

[Zeitsperre]

Looks great. I'll be merging this in after #1650 !

[coveralls]

coverage: 90.193%. first build
when pulling 9a093f9 on fix-agg-sum
into 5db4a75 on master.