forward/backward settings

[emiltin]

ability to set speed and other settings differently for the forward/backward direction.

a specific use is pushing bikes against oneways:
#78

references:
724e96f#commitcomment-2460689

[emiltin]

i pushed updated tests to the feature/opposite_direction branch, and updated the testbot profile, but there seems to be a few problems, so i could not make tests pass. there also seems to be some regresssion on other tests, like "Scenario: Routing on a oneway roundabout # features/testbot/basic.feature:237"

[emiltin]

handling maxspeed:backwards seems somewhat akward, because of they way the backward speed defaults to way.speed if you don't set it.

now that way.speed and way.backward_speed is set separately it seems like way.direction could be eliminated?

we might need to be able to set the mode for each direction, for handling pushing of bikes against oneways.

[emiltin]

would be easier to store (when needed) two plain ways instead of one with backward settings?

[emiltin]

when handling impedance (weight) in lua scripts, this could also be different in the forw/back direction.

[DennisOSRM]

We should have such discussions on the mailing list. The issues get unreadable this way.

[lonvia]

maxspeed:forward and maxspeed:backward currently use a simple 'tonumber' conversion in the lua scripts. Wouldn't it be better to use parseMaxspeed()?

[karme]

i think there is a bug regarding the backward direction (backward_speed might be overwritten by accident)

not a direct fix / but something more like:
https://github.com/karme/Project-OSRM/blob/7b203673d768bc40bd837269185855a888c3577f/profiles/bicycle.lua
might help

[karme]

Emil Tin [email protected] writes:

handling maxspeed:backwards seems somewhat akward, because of they way
the backward speed defaults to way.speed if you don't set it.

now that way.speed and way.backward_speed is set separately it seems
like way.direction could be eliminated?

+1 for that one

but maybe we really should take that to the mailing list?

[DennisOSRM]

@emiltin can we close here?