Enable semantic zooming by default for text labels?

[sehilyi]

If a text mark is mapped to both x and xe channels, we can by default add semantic zooming so that text labels are shown only when it can fit into the region between x and xe:

"visibility": [
  {
    "operation": "less-than",
    "measure": "width",
    "threshold": "|xe-x|",
    "transitionPadding": 30,
    "target": "mark"
  }
]