Bug report: Impossible Nebula

[Amaror]

Hey i just wanted to report a little bug i stumpled upon.
The beacons were so far away from the starting beacon, that it was impossible to reach the ending.
I made a screenshot. It wasn't too bad since it only was the second sector, but it would be quite bad when you on sector 8 or something.

[Matthew]

This has been a rare problem we haven't been able to nail down. If you or anyone else experiences it, please Save & Quit immediately and email me (zasdarq@ftlgame.com) the save file. You can find it in the same location as your profile data, described at this post:

viewtopic.php?f=9&t=1892

[Matthew]

Forgot to say: Sorry about that. Would be very frustrating

[curithwin]

At least you found it in Sector 2 not sector 6.

[Treefrog]

Very similar issue, though in my case cannot reach the quest marker.



Ive included the save file in the attachments.

Hope this bug can get quashed soon enough.

[Steneub]

If I could make a suggestion for a potential fix:

Once the node locations are chosen, run a quick analysis of which nodes are out of reach. If there are any, allow a special case where it is reachable from the closest node, regardless of distance.

[vincentsels]

2012-09-28_000644.jpg

also had it, in sector 4 didn't find this post until after i died (before assuming it was a bug, i thought that i might have missed it so went backwards) so no save game included though :/

[Jotoco]

I have found those as well.

I will remember to save the game when I do it again.

[Matthew]

Also guys, remember that you can turn on pathing from the options as of the new patch. So double check that a sector is indeed impossible! Save games are appreciated for any that you find that are in fact impossible.

[DrNoid]

It seems to me that the problem here is not that a single node is unreachable, but that a cluster of nodes is unreachable. You can't detect that problem by ensuring that every node has at least one reachable other node, since the nodes in an unreachable cluster all have at least one other reachable node, in that same cluster.

The easiest way to detect this situation is to calculate the distance to the starting node, for each node. But since you already seem to be using Dijkstra's algorithm, there must be a bug in either the implementation, or the check to see if all nodes are accounted for...