Factorio Forums

Version: 0.11.19
Platform: Win7 64 bit
Mods: several (uploaded railtest.zip including the save and all the mods and railtest_save.zip including just the save)

When incorrectly connecting curved rail for two connected 90 degree turns (as you would for a basic circle roundabout junction), the game can show false connections where travel isn't currently possible:



The train is attempting to make a right hand turn, which looks like it's possible looking at the smooth turn in the junction, but there's actually no path there. Removing the inner circle makes the broken turn apparent (the curved sections overlap incorrectly), but adding the inner circle hides the break. The game probably shouldn't show a smooth connection where there isn't one in cases like these.

Full image (since it's apparently cut off by the forums): http://i.imgur.com/kN9hKBW.jpg
Save+mods (77MB): https://www.dropbox.com/s/7xhtzl27f3bly ... t.zip?dl=0
Save without mods (9MB): https://www.dropbox.com/s/yhxi7pev3g94b ... e.zip?dl=0

You have a smooth north-south S bend.
And a smooth east-west S bend.

I don't see how the game could detect there is a problem.

Yes, that is not really a bug, just two rails crossing each other in a way, that someone could maybe think, that they are connected.

DaveMcW wrote:You have a smooth north-south S bend.
And a smooth east-west S bend.

I don't see how the game could detect there is a problem.

That's a great way of describing this, and you're right, that's exactly what I got wrong. After doing a bit more research into railroad terminology, it looks like I intended to place a junction, but what I actually placed was a crossing that merely looks very similar to a junction.

I was about to suggest that junctions could be better visually distinguished from crossings than they are currently to avoid confusion, but then I realized that I can just use the in-game signal arrows to tell the difference if there's any confusion; I verified that a signal placed before a junction correctly shows the little yellow arrows branching out on both paths leaving the junction, while the signals before my S-bends-that-look-like-junctions in that picture and save correctly identify that it's really an S bend, not a junction. That's a perfectly acceptable solution; I wouldn't object to this bug being closed. Sorry for the confusion/noise.

I'm sure we will see some more issues around that visualization problem... I hope that gets better, with new track building.