Initial tls phase with custom switching rules

[aminissn]

When I start simulation at a time other than 0, some of the tls initialise in an unwanted phase. I guess sumo calculates the time and tries to guess which phase the tls should be, but then it ignores the condition and next attributes. Not sure if it is an easy fix, but maybe one idea is to start all tls always at phase 0 regardless of simulation begin time? My idea was that the signal stays in a smaller loop with 10 phases unless a but is detected (next=0 in the last phase makes sure the signal doesn't loop over all special phases).

[namdre]

The easiest fix would be to declare an offset attribute that matches the begin time. I could maybe add a symbolic value offset="begin" that simplifies this.