Question for street width is asked endless again and again about streetcomplete HOT 5 CLOSED

I found some time to play around with this and now I believe this is not related to #5545 , but a “simple“ issue with attribute dependencies.

I found it suspicious I could tag the whole street, but not the 2 parts that are a “schikane“ (narrow road for speed reduction). Comparing a working part with a not working part I noticed a difference:

• Road parts having “maxwidth“ are asking repeatedly for maximum width
• Road parts having “width“ are fine

So I changed that not working parts in JOSM and checked what would happen in SC.

Road: https://www.openstreetmap.org/way/1228191464

Initially attributes:

cycleway:left:surface=compacted
cycleway:right:surface=compacted
cycleway=track
highway=residential
lanes=1
lit=yes
maxspeed=30
maxwidth=3
name=Bernadottestraße
sidewalk:left:surface=paving_stones
sidewalk:right:surface=compacted
sidewalk=both
surface=sett
traffic_calming=choker
traffic_sign=DE:274.1[30]
zone:maxspeed=DE:30

Streetview permanently asks: “Wie breit ist die Fahrbahn hier?“ (How wide is the road here?) , to which I reply “3“
—> symbol does not disappear, question is re-asked.

Using JOSM to change “maxwidth=3“ to “width=3“
Upload from JOSM
Refresh in SC

Two new different question appear in SC (Surface type and one-way-road, but those shouldn't matter here) ⟶ problem is gone. Never get asked again.

Then, adding maxwidth=3 again per JOSM and upload: Question does not reappear in SC, issue remains solved

Wild guess: Having maxwidth without width is at least triggering the issue.

Since I now repaired that node: The same issue is here: https://www.openstreetmap.org/way/1228191460 , and I will not touch it.

from streetcomplete.

Thanks for both original report, and for taking extra time for digging into it more @ratti !

I've provided PR which should fix it for next release. (there is linked debug .apk there if you'd like to check how it works now before it's released)

from streetcomplete.

Similar to / possibly the same cause as #5545

from streetcomplete.

Yes, very likely the same as #5545. I believe progress on tracking down the source can only be made after the next update, since I added some logging.

from streetcomplete.

I can confirm that this specific example (w1228191464) is 100% reproducible here in SC v57.1; thus reopening.

Possibly related to ROAD_NARROWERS logic?

from streetcomplete.