Trunk road connects with tertiary glitch

[Grillmannen]

Here a trunk_link connects with a tertiary road: http://www.openstreetmap.org/#map=19/57.77137/14.10716

It looks like the roads don't connect, and that the trunk road continues for a couple of centimeters on the other side of the tertiary. The same phenomenon can be seen for the other trunk_link to the west. I don't think it has been like this before, and at first I thought it was a problem with the data. The data looks fine though.

[matthijsmelissen]

See also #83

The road-width changes might fix this.

[matthijsmelissen]

It seems that the road-width changes do not immediately fix this.

[matthijsmelissen]

We have not really a way to way to fix this without using a flat line-cap, but that would cause issues like in #470. Ideas?

[matthijsmelissen]

I think the only way to solve this is to decrease the difference in size between trunk and tertiary. Would that be an idea?

[Rovastar]

I think we do need to review all road widths (including all links thinner) and add them to all the zoom levels.
Where was that issue I raised ages ago with this. I had a spreadsheet and stuff in it....

[matthijsmelissen]

That review has been done in the meanwhile: #571

It didn't solve this problem, as opposed to for example #83, as here we have a more major road connecting to a minor road. The major road is correctly larger.

[matthijsmelissen]

The current rendering is very ugly, but I don't think we have a way to solve this bug. I will therefore close this issue.

[matthijsmelissen]

To be more precise: we have no way to achieve the following three goals at the same time:

• Have difference in road widths between different highway types of at least twice the casing (currently 1 pixel)
• Avoid gaps in the casing in bended ways (Primary roundabouts rendering glitch #470)
• Avoid having to merge several way segments together in the SQL
• Avoid the current bug