Water areas with tunnel=yes should not render above landcover

[Marc-marc-marc]

layer 1 :
a parking https://www.openstreetmap.org/way/601761935
amenity | parking
layer=1
parking=surface
a road https://www.openstreetmap.org/way/601761936
highway | service

layer=-1
a river https://www.openstreetmap.org/way/601763495
layer=-1
tunnel=yes
waterway=river
the riverbank of this part of the river https://www.openstreetmap.org/way/100709132
layer=-1
tunnel=yes
waterway=riverbank

Expected behavior

show the parking area like all surface parking.
if the tunnel is also rendered, it must be very light and not "overwrite" the render of the parking.

Actual behavior

the riverbank (layer=-1) is render on top of the parking (layer=1) in stead of under it

Links and screenshots illustrating the problem

https://www.openstreetmap.org/way/100709132

[kocio-pl]

It might be solved by #2875 and then removing rendering of underground parkings, see #552 (comment).

[Marc-marc-marc]

this location doesn't have any underground parking.
the parking is at the ground level (a classic parking=surface), the river is in a tunnel under the parking.

[kocio-pl]

Sorry, now I understand. I don't have the idea what to do, does anybody else have some?

BTW: I think this parking should be tagged as layer=0 (or just omit the layer tag).

[kocio-pl]

Maybe we should hide riverbanks when tagged as underground, like #3162?

And maybe the same should be done for natural=water and landuse=reservoir. The common code is here:

openstreetmap-carto/water.mss

Lines 54 to 81 in f8a0f52

	[natural = 'water']::natural,
	[landuse = 'reservoir']::landuse,
	[waterway = 'riverbank']::waterway {
	[zoom >= 0][zoom < 1][way_pixels >= 4],
	[zoom >= 1][zoom < 2][way_pixels >= 16],
	[zoom >= 2][zoom < 8][way_pixels >= 32],
	[zoom >= 8] {
	[int_intermittent = 'no'] {
	polygon-fill: @water-color;
	[way_pixels >= 4] {
	polygon-gamma: 0.75;
	}
	[way_pixels >= 64] {
	polygon-gamma: 0.6;
	}
	}
	[int_intermittent = 'yes'] {
	polygon-pattern-file: url('symbols/intermittent_water.png');
	[way_pixels >= 4] {
	polygon-pattern-gamma: 0.75;
	}
	[way_pixels >= 64] {
	polygon-pattern-gamma: 0.6;
	}
	}
	}
	}
	}

[kocio-pl]

The code for showing tunneled waterways is here - I think areas should be just hidden, but one can use the same code for filtering them:

openstreetmap-carto/water.mss

Lines 120 to 162 in f8a0f52

	.water-lines {
	[waterway = 'canal'][zoom >= 12],
	[waterway = 'river'][zoom >= 12],
	[waterway = 'wadi'][zoom >= 13] {
	// the additional line of land color is used to provide a background for dashed casings
	[int_tunnel = 'yes'] {
	background/line-color: @land-color;
	background/line-width: 2;
	background/line-cap: round;
	background/line-join: round;
	}
	[bridge = 'yes'] {
	[zoom >= 14] {
	bridgecasing/line-color: black;
	bridgecasing/line-join: round;
	bridgecasing/line-width: 6;
	[zoom >= 15] { bridgecasing/line-width: 7; }
	[zoom >= 17] { bridgecasing/line-width: 11; }
	[zoom >= 18] { bridgecasing/line-width: 13; }
	}
	}
	water/line-color: @water-color;
	water/line-width: 2;
	water/line-cap: round;
	water/line-join: round;
	[int_intermittent = 'yes'],
	[waterway = 'wadi'] {
	[bridge = 'yes'][zoom >= 14] {
	bridgefill/line-color: white;
	bridgefill/line-join: round;
	bridgefill/line-width: 4;
	[zoom >= 15] { bridgefill/line-width: 5; }
	[zoom >= 17] { bridgefill/line-width: 9; }
	[zoom >= 18] { bridgefill/line-width: 11; }
	}
	water/line-dasharray: 4,3;
	water/line-cap: butt;
	water/line-join: round;
	water/line-clip: false;
	}

[jaffroy62]

This could be a similar symptom, if the bridge is removed the riverbank is rendered on top of the parking area.

Expected would be the that the amenity=parking area is painted on top of the bridge.

Assumption: could it be that the layer tag of the parking area is not interpreted correctly?

https://www.openstreetmap.org/#map=19/48.52899/9.06783

Layers in this case:

bridge: layer=1
riverbank: layer=0 (standard)
parking area: layer=1

[kocio-pl]

Wild guess: probably this is because layer tag is not global in this style, it works only in the internal "data layer" (for example it within roads, but parking is in amenity layer).

[jaffroy62]

I observed another unclear appearance: Here a bridge goes over a residental highway. In my opinion the residental in this example should be painted with shaded appearance.

https://www.openstreetmap.org/#map=19/48.53115/9.03813

btw. the parking area underneath the bridge is rendered correctly :-)

[matkoniecz]

Here a bridge goes over a residental highway

See #1799 for more discussion about road/bridge layering.

[HolgerJeromin]

In my opinion the residental in this example should be painted with shaded appearance

Just add covered=yes and you have what you want (and did not lie for the renderer).

[jeisenbe]

We could try not rendering water areas, including waterway=riverbank and natural=water, when tagged with tunnel=yes. The waterway line should still show as a tunnel.

[jaffroy62]

Does this mean that the bridge problems presented in my above posts have to be moved to an seperate issue?