Box warp doesn't treat subpaths as unified object

It would make sense that a box warp on a group of paths would treat the group as a single object and keep all subpaths sized and positioned relative to each other. In practice, this operation works on each subpath as separate objects, which isn't very useful.

I'm not sure conceptually where you should be able to distinguish between the two operations, but performing it on a layer containing multiple subpaths seems like the trivial scenario when the earlier behavior would be desired. Maybe it's as simple as a boolean input e.g. "Combined" on the box warp node.

No files, tests, or entry points are named. Start by reproducing box warp on a layer containing multiple subpaths, then determine how the operation should distinguish unified treatment from separate treatment; done means the selected behavior preserves the subpaths' relative size and position.

Vom Indexierungsmodell aus dem Issue-Text verfasst.