trains prefer to stay in lane at all intersections

[kianzarrin]

fixes #1475
TMPE.zip

• I check if the lane is a track only lane. that is train/metro/monorail/tram but mixed tram/car does not count.
• I also check if both source/target segments have equal number of tracks.
• if both of the above conditions are met then : Routing manager only creates one lane transition which is between the two lanes with equal similar Lane Index.
• user can override the above behavior using lane connections

PS: Created TrackUtils class even though it has only one function. I might add more stuff to it in future if I deal with Continues junction barrier stuff (#1223) or if I bring Hide-Unconnectred-Tracks to TMPE. (maybe I should call it DirectConnectUtils)
PS2: I used $"{varaible}" to log stuff because it was more reliable. I hope you don't mind: #1484

EDIT: for bidirectional/station tracks I fall back to original code.

[kianzarrin]

Test ✔️

[kianzarrin]

Test: lane connection tool can be used to change lane connections ✔️

[kvakvs]

Code looking fine

[kianzarrin]

This logic fails for bi-directional tracks. it actually makes thing worse:

I need to reverse lane transitions for bi-directional tracks.

The rest of the code also use similar lane index and to calculate distance or something and does not care about bi-directional tracks problem.

I suspect if you put bidirectional roads in reverse direction trains will prefer to criss-cross even in the original code.

[kianzarrin]

I will ignore (i.e. fall back to old code) bi-directional lanes in this PR. I created #1486 to cover bi-directional problems for all of routing manager (note that the code already had problem with bidirectional lanes even before this PR).

[kianzarrin]

Its working as expected now: all lane transitions are created for bi-directional tracks

[originalfoo]

will this PR benefit from some of the extensions you added in #1487

[kianzarrin]

@aubergine10

will this PR benefit from some of the extensions you added in #1487

No. All I need here is to ignore train/bidirectional lanes which I can detect using IsBidirectional(). I have deleted that in the next PR because its just a workaround for work that is outside of scope of this PR (#1483 (comment)).

In any case if you think the code is good you can approve this PR.