Block stairs are not following the curve

[axrolld]

What happened?

With allowances :

Without :

What did you expect to happen?

The blocks are supposed to follow the green curve if there's one, the blue curve if not.

How can we reproduce it (as minimally and precisely as possible)?

This does not always happen but i had this bug on train 876401 here.

What operating system, browser and environment are you using?

• Browser: Chrome
• OS: W11
• Env: internal staging

OSRD version (top right corner Account button > Informations)

3a2b863

[shenriotpro]

From @Khoyo : we should probably try to reproduce locally, so we can attach a debugger and check if it's a standalone simulation issue, or related to signals.

[shenriotpro]

While reproducing with @axrolld we noticed 2 things:

1. trains imported from private data don't seem to sync well with the editor, e.g. for stops durations (should probably be a separate issue)
2. the (blue) curve actually seems to be the issue: it apparently ignores stops durations (effectively considers 0)

The good news from 2. is that we can reproduce easily by creating a train and adding a (non-null) stop

[shenriotpro]

One issue is that we "miss" the first stop, I put logs below for an example on small_infra where the position and stop position are never equal. /~https://github.com/osrd-project/osrd/blob/2a2abf970bd562137ce7ff3389ae5fa83219d50f/core/src/main/java/fr/sncf/osrd/standalone_sim/EnvelopeStopWrapper.java#L57
Note that in this case the first stop has duration 0 (station is not a "via").

2024-01-30 11:08:37 623.0183612728907
2024-01-30 11:08:37 653.636
2024-01-30 11:08:37 false
2024-01-30 11:08:37 667.2696211639147
2024-01-30 11:08:37 653.636
2024-01-30 11:08:37 false

Checking all stops (instead of just the next one) could fix the issue, but the problem seems deeper.

[shenriotpro]

From @eckter: we should probably check we are passed the stop, instead of at the stop.