Kevin Kreiser

it's also very useful if you're worried about buildings coverage in your map display since it has a bunch of added sources for that

yeah seems like something is systemically wrong.. i wonder if we are suffering from something more exotic like the initialization fiasco: https://en.cppreference.com/w/cpp/language/siof i dont think so because we always use...

this https://github.com/valhalla/valhalla/commit/0f6367af7b2c0f974f22340a497e34ee601b340a#diff-87881ad0664f8b3ec64493ebbc31fd0b677428c47f75197fc8de2d5e7cacbe01R61 added a whopping 40 bytes to the node. previous osmnode was only 17 bytes, now it is 27, a 60% increase. osmnode is used in all way_nodes even...

when we hook up gtfs we look for edges to connect that have pedestrian access and have a "valid" use as defined here: https://github.com/valhalla/valhalla/blob/efff65d35c68c9198d504e07087d3ae2b71e5e1f/src/mjolnir/transitbuilder.cc#L34-L37 note that `kPlatform` is not one...

yeah just to spell it out here as well, `ignore_restrictions` is meant to ignore any restrictions. if you are using it and expect it to respect certain restrictions then you...

sorry for being quiet so long. the whole point of the expansion action **is** debugging. i know that over the years others have found other interesting use cases for it...

i thought it was safe for c++ to put whatever bytes it wants in protobuf strings. protobuf actually checking the encoding of the strings when its parsing, what a waste...

looking at the protobuf source it seems its been there for at least 10 years... that sucks

changing to bytes is a technically a breaking change for the pbf api which we have supported for some years now. on the wire, bytes and string are the same,...

yep you are right its a product of hysterical raisins! in the early days we didnt recover shortcuts before narrating the path i added that once we started supporting live...