[devel-new a6879f1a08] No route to waypoint

[MCollett]

Waypoints seem to bamboozle the routefinder. 

A toy example:
http://simutrans-germany.com/files/upload/WaypointTest2.sve
The stagecoach at Barrow Town Hall cannot find the waypoint that is next on its schedule, even though it is just around the corner.

[jamespetts]

Sorry for missing this: I did not receive the notification e-mail for some reason. Non-reversing waypoints are routed through to the next waypoint beyond, so check to see that that next stop is reachable. Stage coaches have a range limit in the latest development code, so check to see whether this is exceeded, too (waypoints do not count: the range limit is measured between stops).

[MCollett]

Sorry for missing this: I did not receive the notification e-mail for some reason.

Remember that if more than one post is made to a board, you only get notification for the first.  Perhaps you also missed this one?

Non-reversing waypoints are routed through to the next waypoint beyond, so check to see that that next stop is reachable. Stage coaches have a range limit in the latest development code, so check to see whether this is exceeded, too (waypoints do not count: the range limit is measured between stops).

In the example I posted, the coaches were happily shuttling between Sowerby and Barrow before the waypoint was added, so neither of those can be the reason.  (And since the waypoint has yet to be successfully visited, presumably it is by default still being treated as reversing.)

[jamespetts]

This problem was limited to vehicles with a limited range and is now fixed. Thank you for the report!

[MCollett]

This problem was limited to vehicles with a limited range and is now fixed.

Running the example game with the newest version, the currently stuck vehicle happily finds a route to the waypoint.  Unfortunately, when the next scheduled departure time comes round, the game crashes with:

FATAL ERROR: vector_tpl<T>::[] - 7koord3d: index out of bounds: 1 not in 0..4294967295
Aborting program execution ...

(a puzzling message, since 1 certainly is somewhere between 0 and 2^32).

[jamespetts]

I think that I have fixed this - can you have another go?

[MCollett]

I think that I have fixed this - can you have another go?

Looks OK now.  The industry connection reload problem also seems to be fixed, so I think the only significant (non-cosmetic) issue remaining from the bugs I've reported in the last few days is the misreporting of industry capacity.

Thanks,
Matthew

[jamespetts]

Splendid, thank you for checking. Glad that these things are now working.