It doesn't seem to, I've checked that before myself.
Connector jumps, plant corrupt
I was hit by this bugger again and I decided to trace this in the decompiled 0.8.7b/b615 code. It looks like blueprint preview and actual placement use different transport endpoint-joining rules.
StaticTransportPreview knows whether each transport endpoint connects to something inside the blueprint. It passes that information as disallowMiniZipAtStart/End during preview validation.
During actual placement, the config-based path through TransportsCommandsProcessor.TryBuildTransportFromConfig does not appear to carry those endpoint flags forward. Transport validation is therefore performed with both flags set to false.
If a copied endpoint overlaps an existing transport, commit-time validation may treat the endpoint as eligible for a new join. TransportsConstructionHelper.CanJoinTo then cuts the overlapping transport and creates a mini-connector there. Visually, this looks like the blueprint connector moved by one tile.
As a test, I forced both endpoint-protection flags to true during config-based transport placement. The pasted result then matched the preview exactly.
Possible fixes:
- Carry the preview’s per-end connection state through the placement command and use it during commit.
- Alternatively, determine on the simulation side which transport endpoints connect within the submitted batch, and suppress automatic joins only at those endpoints.
I may be missing some intended behavior here. This probably isn’t risk-free: partial blueprints, connections to transports outside the blueprint, replacement order, undo/redo, and multiplayer determinism may all need consideration. Globally protecting both endpoints worked as a diagnostic, but is likely too broad for a production fix.
Marked 1 topic as a duplicate of this one: