don't ask about relation membership handling when combining 2 ways which are part of the same multipolygon

[dieterdreist]

sometimes when you combine 2 ways which are both part of a MP relation, you have to manually confirm that you want to keep one and remove the other. I did not find out when this happens, I have had it ocassionally, and it did not seem necessary because of unambiguity.

[skyper]

An example and the status report would be helpful.
As far as I understand, JOSM is asking for user decisions for relations which do not include all the ways as members with the same role.
In my eyes, this is useful.
The only situation, I am annoyed of it is when using Replace Geometry with all memberships on the new way. In this situation an option to keep all would be helpful as it is no fun to click twice for each membership.

[GerdP]

This happens when the way members are not sorted, e.g one is in position 2 and the other in 5. I think it is correct to ask in this case.

[dieterdreist]

Thank you GerdP, this was it, it happens when they are not sorted. IMHO as long as it is unambiguous (all ways that are combined are part of the same relation and not of others, maybe also tag comparison), it should not matter whether they are sorted.
The actual warning message is very strong:
"Conflicts when combining ways - combined way is '0 (5 nodes)" is the title. I do not see any conflict at all in this situation.

Then there is a text which says "the default is to keep the first way and remove the other ways that are members of the same relation: the combined way will take the place of the original way in the relation".

If this is the default, why isn't it preselected?

On a sidenote, while performing some tests, something really strange happened: made a multipolygon consisting of several outer ways and combined them, and on purpose decided to keep all results in the relation. As a result, the relation editor shows something crazy (2 closed outlines where one is basically 1 way going back and forth), i.e. the thing looks (almost) valid in the relation editor and validation does not see any problem. Please see attached screenshots and data.

[dieterdreist]

invalid MP resulting from combining members and selecting wrong options, but which is not caught by validator and seems valid in the MP editor

[dieterdreist]

dialogue before combining

[dieterdreist]

after combining

[skyper]

I get an error warning, Multipolygon member repeated with same role, with you example.