Fixed #9406: By treating aerialway=station as valid start/end point which suppresses the false warning

[Razen04]

Description

Currently, iD warns for paths which are not connected to any other route any is contained within a building like gondola stations which is a false positive. This has been explained in detail in Issue #9406.

Fix

The fix to this is simple, we just add an extra check in the isConnectedVertex function to check if a vertex in the path has aerialway=station and return true for that, which means the path is valid to have start/end in the gandola station even if they are not connected to any other roads or highway(which they aren't supposed to).

Fixes #9406

Before:

After the if statement:

References

This is similar to issue #3906 - which showed problems with foothpaths connecting two buildings.

[k-yle]

we have another preset, called Aerialway Stopping Location, which seems to be used in much the same way, and also seems to be a routing sink.

Should this logic apply to that tag as well?

[Razen04]

I think public_transport=stop_position && aerialway=* should be treated the same way as aerialway=station.

In both cases, the mapper is explicitly marking a boundary where pedestrian movement ends and aerial transport begins.

This is different from public_transport=stop_position for buses, trams, or trains, where pedestrian connectivity is expected to continue.

So should I add that case too in the code?

[matkoniecz]

I think so, but I would be also fine with doing it as a next PR (note: not a maintainer)

[Razen04]

Let's see why k-yle says. The fix is just one line more i suppose.

[matkoniecz]

@k-yle do you have by any chance time to look into that?

[k-yle]

it should be a trivial change, i would fix it in this PR imo, so that we have a centralised place to this discuss this topic

[Razen04]

I'll do it.

[matkoniecz]

appears to be done, from what I see

[Razen04]

I messed it up in between, lol. Anyways I think I made. the correct change, do look into it once.

[matkoniecz]

Yeah, this 579 changed files looked a bit suspect :)

[matkoniecz]

I guess it would also count say public_transport=stop_position aerialway=no ?

And public_transport=stop_position aerialway=whatever_gibberish ?

Should requiring specifically aerialway=station or public_transport=stop_position aerialway=yes be needed?

[tyrasd]

I fixed this slightly differently in cd255c8: the aerialway ways themselves can be considered to be part of the routable network in such cases. There's no need to do the hacky treat vertex as routing source for this particular issue.