incorrect warning on roundabouts: "street with odd number of lanes... and no oneway"

[verdyp@…]

What steps will reproduce the problem?

1. enable validator and check lanes
2. load way 4765121 (roundabout with 3 lanes)
3. validate

What is the expected result?

No error.

What happens instead?

Complains with "missing tag - street with odd number of lanes but without lanes:forward and lanes:backward and no oneway".

This rule is incorrect because roundabouts are *implicitly* oneway (so there's no explicit oneway=yes as expected by this rule).

That way is a correctly closed circle (anticlockwise as expectd in France)
It is tagged with:

• highway=primary
• junction=roundabout
• lanes=3
• name=Rond-Point de Bray

Roundabouts with odd (1 or 3) lanes are common. This is definitely not an error.

Please modify your rule wo that "oneway=*" becomes "oneway=* OR junction=roundabout" (possibly other highways may be implicitly oneway)

Please provide any additional information below. Attach a screenshot if possible.

URL:http://josm.openstreetmap.de/svn/trunk
Repository:UUID: 0c6e7542-c601-0410-84e7-c038aed88b3b
Last:Changed Date: 2017-02-02 01:22:52 +0100 (Thu, 02 Feb 2017)
Build-Date:2017-02-02 02:33:28
Revision:11526
Relative:URL: ^/trunk

Identification: JOSM/1.5 (11526 fr) Windows 10 64-Bit
Memory Usage: 2684 MB / 5337 MB (1572 MB allocated, but free)
Java version: 1.8.0_121-b13, Oracle Corporation, Java HotSpot(TM) 64-Bit Server VM
Screen: \Display0 1920x1080, \Display1 1080x1920
Maximum Screen Size: 1920x1920
VM arguments: [-Djava.security.manager, -Djava.security.policy=file:<java.home>\lib\security\javaws.policy, -DtrustProxy=true, -Djnlpx.home=<java.home>\bin, -Djnlpx.origFilenameArg=D:\Utilisateurs\<user.name>\Downloads\josm.jnlp, -Djnlpx.remove=true, -Djava.util.Arrays.useLegacyMergeSort=true, -Djnlpx.splashport=51811, -Djnlpx.jvm=<java.home>\bin\javaw.exe]
Dataset consistency test: No problems found

Plugins:
+ Mapillary (v1.3.2)
+ apache-commons (32994)
+ apache-http (32699)
+ ejml (32680)
+ geojson (52)
+ geotools (33042)
+ jts (32699)

Tagging presets:
+ https://raw.githubusercontent.com/yopaseopor/traffic_signs_preset_JOSM/master/FR.zip

Last errors/warnings:
- W: Unable to get color from '' for color preference 'extrude.main.line'
- W: Unable to get color from '' for color preference 'improve.way.accuracy.helper.line'
- W: Unable to get color from '' for color preference 'make.parallel.helper.line'

[anonymous]

Note: unlike bidirectional highways using lanes=* (that must have even value for lanes=*, assuming that half of them is forward and the others are backward, or otherwise must specify lanes:forward=* and lanes:backward=*), all roundabouts are implicitly oneway, and lanes are necessarily forward (or all backward if "oneway=-1", but this should be avoided by reversing it and dropping the superfluous "oneway=yes").
Roundabouts most frequently have 2 ways (except mini-roundabouts that have generally only 1).

But larger roundabouts 3 ways are also common, notably those used by regular bus lines or connecting primary highways and avenues: the inner way is for turning left, the median way for going forward, the outer way for turning right. There are also some large roundabouts with 5 ways (the 3 central ways may be used for forward, the two inner ones may be used for turning left or could be used by long vehicles.)

[anonymous]

Also additional ways (still forward by default) may be reserved for a cycleway, or a busway (possibly with a bus stop if it's the outermost lane).

We should not need to explicitly set ":forward" on "lanes" tags for any roundabout

But there may exist sometimes situations with an outer ring going backward or being directional (for pedestrian or for bicycles), and explicitly set as "backward" (substracted from the total number of "lanes=*"), but usually there's a separator and these supplementary lanes should be traced as a separate concentric highway (not tagged as a junction=roundabout).

[verdyp@…]

Roundabout with 3 lanes

[Don-vip]

Closed as duplicate of #14323.

[anonymous]

Closed as duplicate ? But #14323 is closed for "being solved". And it is not!

[Klumbumbus]

Replying to anonym:

Closed as duplicate ? But #14323 is closed for "being solved". And it is not!

It was fixed in version r11529 and it no longer warns in josm-latest. If you use the tested version of josm, then you have to wait a bit until a new tested version will be released (which is sheduled at the end of every month).