Opened 13 months ago
Last modified 13 months ago
#23966 new defect
MapCSS validator: Turn:lanes:forward and change:lanes:forward with oneway=* tag is marked as a suspicious combination
| Reported by: | Owned by: | team | |
|---|---|---|---|
| Priority: | normal | Milestone: | |
| Component: | Core validator | Version: | |
| Keywords: | Cc: |
Description
https://josm.openstreetmap.de/browser/josm/trunk/resources/data/validator/combinations.mapcss#L316
I can definetly say that turn:lanes:forward and change:lanes:forward are exactly the same as turn:lanes and change:lanes on one-way roads and can be interchanged.
it only becomes a problem when its turn:lanes:backwards and change:lanes:backwards combined with oneway=yes.
I do agree that it is a totally unnecessary suffix, but this in itself should not provoke a suspicious tag.
this list is directly used in osmose.
Attachments (0)
Change History (3)
comment:1 by , 13 months ago
comment:2 by , 13 months ago
@taylor.smock so your reasoning being "so we can know the tag could be outdated", that is not the job of the validator.
maybe it could be marked as info instead.
comment:3 by , 13 months ago
I think that marking it info would defeat the purpose, since in the vast majority of cases it does point to a true suspicious situation:
- Either it exists already, in which case there's a large probability of it being an unintended leftover
- Or the user isn't aware of best tagging practices
- Or the user is under the impression that (s)he is adjusting a two-way road instead.
Making it info-level would hide it from most users (or have them ignore it), thus likely letting cases slip through.
So I'm with Taylor
Like with all QA, there'll always be exceptions and it doesn't necessarily point to an error, it can also just be a suggestion for/pointer to improvement.



No; I think it should be suspicious. For example, I've had to remap a bidirectional road that was turned into a divided highway. It can be easy to miss a section where there were
*:lanes:forwardtags when updating the road.IMO, a stretch of road with
oneway=yes+ (turn:lanes:forward|change:lanes:forward) should have the turn/change lanes checked since they are likely left over from a conversion of the road from a bidirectional road into a divided highway or just made a oneway with no modifications (example: two roads surrounding a block are made oneways in opposite directions).