When the step_count is zero because the step_size is less than the length of the lane then it generates a lane_position at the beginning of the lane, in contrast with the old behavior where none lane position was generated in that case, occasioning an error downstream.
Why did we miss it in bionic?
delphyne_city demo was working on bionic(python 3.6.9) because the random lanes that were being selected were always the same (because the seed was the same) and it happened that, luckily, none lane without available lane_position was selected. When running in focal(python 3.8.10), the selected random lanes changed and now some lanes, which no lane_position were generated, were now selected.
resolves https://github.com/ToyotaResearchInstitute/delphyne_demos/issues/30
When the
step_countis zero because thestep_sizeis less than the length of thelanethen it generates alane_positionat the beginning of the lane, in contrast with the old behavior where none lane position was generated in that case, occasioning an error downstream.Why did we miss it in bionic?
delphyne_citydemo was working onbionic(python 3.6.9) because the random lanes that were being selected were always the same (because the seed was the same) and it happened that, luckily, none lane without availablelane_positionwas selected. When running infocal(python 3.8.10), the selected random lanes changed and now some lanes, which nolane_positionwere generated, were now selected.