When deduplicating nodes, give preference to existing nodes, instead of new nodes

[fsteggink]

For importing landuse data in both Canada and the Netherlands it is necessary to deduplicate nodes. Thankfully this is happening fast now, due to an earlier ticket being resolved :) However, in case of the Canadian import I would like to reuse nodes of forest areas I've uploaded earlier, but the ways / relations can be removed, because the attributes have been updated.

When I perform a node deduplication with the validator, it often gives preference to new nodes, instead of existing nodes. I hope that this can be changed be changed. Reason: much less upload of data will be required. Currently I'm importing a sheet near Trois-Pistoles, Québec, and there are about 50k changes (including other data). Of this about 32k are new features, and 16k are features to be removed. Many of the latter features are "old" nodes which are being uploaded again, in exactly the same position.

[bastiK]

Cannot reproduce, please update to latest version and provide a small test file where this happens (and note the node id).

[fsteggink]

I'm not sure how I initialized the deduplication, but if I perform it before I remove redundant features (from an earlier import by myself), things seem to work out as I hoped. I'll close this ticket.