Finally finally we're wrapping up #136. The code is hopefully easier to understand now, and the results are slowly improving. The order updates and transformations happen is still sensitive and we'll have to do something more rigorous in #159.
Before this PR, I got rid of shrinking roads in #167 and updated one test's OSM data in 2b89cc120c45940d5a2b3e861c598d01cc78b598. This reduced the diff from this PR and made it easier to understand the subtle ordering between things.
A sample of changes. Nothing gets much worse; lots of places change in kind of neutral ways. Highlights: kingsway_junction:
quad_intersection
seattle_slip_lane:
tempe_light_rail:
Actually going to not merge this one until a review, since it's pretty important
Finally finally we're wrapping up #136. The code is hopefully easier to understand now, and the results are slowly improving. The order updates and transformations happen is still sensitive and we'll have to do something more rigorous in #159.
Before this PR, I got rid of shrinking roads in #167 and updated one test's OSM data in 2b89cc120c45940d5a2b3e861c598d01cc78b598. This reduced the diff from this PR and made it easier to understand the subtle ordering between things.
A sample of changes. Nothing gets much worse; lots of places change in kind of neutral ways. Highlights:
kingsway_junction
:quad_intersection
seattle_slip_lane
:tempe_light_rail
:Actually going to not merge this one until a review, since it's pretty important