There were 3 manual steps here:
1) Tag sidewalk=left,right,both along many one-way streets in OSM where
the inferred value was very wrong, and causing a large difference
from smp_sidewalk_widening
2) Clear out smp_sidewalk_basemap, broken by OSM ID changes. This has
never actually (and still does not) contain real baseline sidewalk
widths, but still might eventually.
3) Manually update smp_sidewalk_widening, broken by OSM ID changes.
The simulation results with the proposed edits now don't cause as many
spurious route changes.
PolyLine instead of list of points.
This exposed two edge cases, fixed here:
1) While merging short roads near junctions, we sometimes create a loop.
Filter it out immeditely.
2) If an OSM way is physically sharing geometry with another one
invalidly, detect that and discard it much earlier.
The main reason is to unblock osm2lanes cutover by avoiding
https://github.com/a-b-street/osm2lanes/issues/153. A secondary reason
is to pull in fixes in SLU for #818.
But of course grabbing fresh data wasn't straightforward:
- Need to collapse small junction=circular ways just like roundabouts,
or else Montlake starts gridlocking. And so actually, regenerating all
maps, in case any happen to use this.
- Had to update a bunch of proposals since OSM IDs changed. I just used
the automatic repairs with some edits filtered out. Fixing edits
really is quite tedious when IDs change; we need something more
robust. (Just give up, describe geometry, and always snap to the best
match?)
- Had to update the one baked-in traffic signal for similar OSM ID
reasons, with similar hassle.
We need this because we have a data source for SMP that we can use with
extra attribution, but not upload to OSM.
This whole commit is gross hacks; going forward, we need a proper
process for upstreaming stuff in OSM.
by chance in #870.
A vehicle exits a driveway and blocks a lane on its way out. When we try
to clear the static blockage, the queue calculation recurses to resolve
distances on the queue where that vehicle is entering. But since we
temporarily don't have the car in the list of cars (for the borrow
checker), it was crashing.
One of those "2 hours to figure out, 30 seconds to fix" bugs.
Back in #819, I broke things by creating the sidewalk+transit graph
before transit routes had been set up. This causes immediate crashing
when editing the 2 maps with transit (Arboretum and SF).
And clean up some other things that RDP does better.
Fallout from regenerating everything:
- Enfield borough crashed, so removed it
- All UK scenarios are now much bigger, due to the changes in #853 being
picked up
- Poundbury gridlocks now due to that
This was happening when trying to draw pedestrians sometimes.
Regenerate everything. Some maps/scenarios affected, but in tiny
imperceptible ways. No screenshot diff.
