Unable to combine legends when using line mark with "point": true or separate line and point layers

[labkey-alan]

Bug Description

Repro for the bug can be found on this vega editor url.

My spec, minus the data, looks like:

{
  "$schema": "https://vega.github.io/schema/vega-lite/v6.json",
  "data": {
    "values": []
  },
  "mark": {"type": "line", "point": true},
  "encoding": {
    "x": {
      "field": "maxSpeed.value",
      "type": "quantitative",
      "title": "Max Speed"
    },
    "y": {
      "field": "weight.value",
      "type": "quantitative",
      "title": "Weight"
    },
    "color": {
      "field": "type.value",
      "type": "nominal",
      "title": "Type",
      "legend": {"title": "Type", "type": "symbol"}
    },
    "shape": {
      "field": "type.value",
      "type": "nominal",
      "title": "Type",
      "legend": {"title": "Type"}
    }
  }
}

With the above spec I would expect the legend to have one section "Type" and to include one legend value per type, with the color and shape represented per value on the legend:

However, what I get are two distinct legends, one for the shape, and one for the color. If I change the mark to use point like so:

"mark": { "type": "point" },

I get a properly combined legend:

Checklist

• I checked for duplicate issues.

[labkey-alan]

Here is another visualization that I created that shows the same problem, but with a slightly different Vega-Lite spec. Here is the spec without the data (the link to the Vega Editor includes the data):

{
  "$schema": "https://vega.github.io/schema/vega-lite/v6.json",
  "data": {
    "values": []
  },
  "layer": [
    {
      "mark": {"type": "line", "point": false},
      "encoding": {
        "color": {"field": "type.value", "type": "nominal", "title": "Type"}
      }
    },
    {
      "mark": {"type": "point", "opacity": 1, "filled": true, "size": 55},
      "encoding": {
        "color": {"field": "type.value", "type": "nominal", "title": "Type"},
        "shape": {"field": "type.value", "type": "nominal", "title": "Type"}
      }
    }
  ],
  "encoding": {
    "x": {
      "field": "maxSpeed.value",
      "type": "quantitative",
      "title": "Max Speed"
    },
    "y": {"field": "weight.value", "type": "quantitative", "title": "Weight"}
  }
}

So, instead of trying to use just the line mark with "point": true I went ahead and created two individual layers, one for the lines, one for the points. Each layer defines their own color encoding, and the point layer includes an encoding for shape. I would expect this to result in a combined legend, but it does not. The results are basically identical to using only the line mark with "point": true. I tried disabling the legend for the line layer but that just drops the color values from the legend.

[ByzantineFailure]

Noting that we encounter this issue as well when attempting to use a shape mark to apply symbols to different series in a stacked bar chart for accessibility purposes (need a non-color way to differentiate them).

{
  "$schema": "https://vega.github.io/schema/vega-lite/v5.json",
  "description": "A simple bar chart with embedded data and shapes in the center",
  "data": {
    "values": [...],
  },
  "transform": [
    {
      "stack": "b",
      "groupby": ["a"],
      "as": ["y0", "y1"]
    },
    {"calculate": "(datum.y0 + datum.y1) / 2", "as": "midpoint"}
  ],
  "encoding": {
    "x": {"field": "a", "title": "Category", "type": "nominal", "axis": {"labelAngle": 0}},
    "color": {
      "field": "c",
      "title": "Product",
      "type": "nominal",
      "scale": {"scheme": "category20"}
    },
    "shape": {
      "field": "c",
      "type": "nominal"
    }
  },
  "layer": [
    {
      "mark": {"type": "bar"},
      "encoding": {
        "y": {"field": "y0", "type": "quantitative", "title": "Value" },
        "y2": {"field": "y1", "title": "Value", "type": "quantitative"},
      }
    },
    {
      "mark": {"type": "point", "filled": true, "stroke": "black", "size": 100},
      "encoding": {
        "y": {"field": "midpoint", "type": "quantitative"},
      }
    }
  ]
}