Trains entirely ignoring signals

[wlindley]

For the past month or so, builds from latest source all result in trains utterly ignoring signals and every train on the map crashing into each other. This occurs in fresh new games (1850 start, for example) as well as all my old saved games.  I have been building signalled lines for many years; did something fundamental change about how we have to signal lines, or is the codebase broken? Very discouraging.

[jamespetts]

For the past month or so, builds from latest source all result in trains utterly ignoring signals and every train on the map crashing into each other. This occurs in fresh new games (1850 start, for example) as well as all my old saved games.  I have been building signalled lines for many years; did something fundamental change about how we have to signal lines, or is the codebase broken? Very discouraging.

I cannot reproduce this, I am afraid: using the demo.sve saved game, signals work correctly. Can I check whether you have tested with the downloaded binaries or whether you are self-compiling?

[wlindley]

Self-compiled, but directly against the latest sources.  Here is an example save file.

Two trains are leaving Bumbleham. The second should wait at the signal ("time interval with telegraph" style) but instead proceeds right through it and crashes into the first at the other station.

Edit: Moving time forward and replacing the signals with "Semaphore Stop" makes things work as expected.  Did the function of the time-with-telegraph signals change? And if so, how is it even possible to arrange time-with-telegraph to work for a simple line like this? It always worked before just as shown. Very confused.

[wlindley]

Confirm the save file works the same in the current nightly binary with the current nightly pakset.  Trains blithely pass signals at danger and collide into each other.

[jamespetts]

Thank you for the report. I am currently doing some intensive work on the 15.x branch, but will look at this when I next turn back to the master branch for bug fixes and the incorporation of patches.

[jamespetts]

Apologies for the delay - I believe that I have now fixed this. I should be grateful if you could re-test with the next nightly build.

[wlindley]

Yes, the latest nightly does resolve the problem. Thank you!  (Now for me getting my local git repository to properly build things again... sigh)