Open matkoniecz opened 10 years ago
There are performance problems involved with solving this. We could render oneway arrows immediately on top of the road it belongs to, but it would imply having three times as much road rules in the xml (for oneway=yes, -1 and no).
oneway=yes, -1 and no
Given the sparsity of oneway values other than yes, we could normalize all oneway values to null or yes by reversing -1 and turning no into null without any real performance issues. This would also help make the MSS simpler by removing the -1 case
Still it would mean duplicating the number of rules for roads.
Is the number of rules a significant performance issue? I've found the number of layers generally more significant. I would like to take a look at reversing oneway=-1
ways to only get oneway=yes
or oneway IS NULL
out of the layers
from #953 http://www.openstreetmap.org/#map=17/59.30845/15.41528
motorway_link arrow renders over the secondary road
(fixing #284 would fix also this case)
And sometimes results are really misleading - see http://www.openstreetmap.org/#map=19/50.07165/19.94518
Slightly different manifestation of this issue. Example where oneway arrows get rendered outside of the typical visual context of a road entirely.
The arrows jump around at different zoom levels which without the visual context of the parent road element, due to being underground, is especially confusing until you realize what's happening as they seem like they're meaningfully pointing and moving to draw attention as you zoom.
(I notice the label "PATH" (by The Archer) for an underground walkway is also rendering seemingly orphaned. Similar or different issue?)
Also, would it make sense to tweak the issue title? I initially thought this was going to be related to roads tagged covered=yes such as building passages.
Also, would it make sense to tweak the issue title? I initially thought this was going to be related to roads tagged covered=yes such as building passages.
Can you propose a better one?
@cody-somerville - that is a different problem for which we already have #529.
By the way this is also closely related to #2213 and #4401.
http://www.openstreetmap.org/?mlat=50.05523&mlon=19.94447#map=18/50.05523/19.94447
Residential is not oneway, private service roads rendered below are.
Related to #597 #681 #245
Fixing #284 would probably also fix most occurrences of this bug.