META: CopyNodesRecursively vs Brain

[mhsdesign]

Except the big known issues with CopyNodesRecursively

• 1.) BUG: CopyNodesRecursively vs Tethered nodes
• 2.) BUG: CopyNodesRecursively vs Content Dimensions

I think that once you start thinking about it, it does not stop seeing problem with it :D

• The NodeSubtreeSnapshot is NOT validated and taken for correct, imo it must not be possible for a user to do that accidentally with CopyNodesRecursively.

• On Rebase there is NO validation of the $childNodes to insert, we just turn the serialised CopyNodesRecursively directly back into NodeAggregateWithNodeWasCreated events, defeating the sense of replaying commands partially as we kinda copy events here:

  • This is problematic because the conditions when a node tree snapshot was created might have changes:
    • The target node that was referenced by a child node might be deleted,
    • A child nodes node type might have been adjusted and the new property schema is not validated
    • ... more?

• When copying nodes they will lose information about their subtreetags

[mhsdesign]

In todays weekly Christian, Denny, Paula, Basti and me discussed that we should probably have a Serialised version of the command instead of adding the NodeSubtreeSnapshot to the API one.

Instead, the api one would contain:

NodeAggregateId $startNodeAggregateIdToCopy,
PropertyValuesToWrite $additionalPropertyValuesToWrite,

And we probably need to keep the NodeAggregateIdMapping on the api one for testing.