MapCSS classes are not set if defined in information-level rules

[mkoniecz]

What steps will reproduce the problem?

1. create highway=path/footway
2. add tourism=information, information=guidepost node on the way
3. run validator

What is the expected result?

Validator is not complaining, guideposts (especially in mountains) may be placed directly on a path

What happens instead?

Validator complains - tourism node connected to a highway (1)

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

In attachment - one of many potential examples, found via ​https://duckduckgo.com/?q=drogowskaz+tatry&t=canonical&iax=1&ia=images

URL:http://josm.openstreetmap.de/svn/trunk
Repository:UUID: 0c6e7542-c601-0410-84e7-c038aed88b3b
Last:Changed Date: 2017-07-07 22:32:46 +0200 (Fri, 07 Jul 2017)
Build-Date:2017-07-08 01:32:27
Revision:12456
Relative:URL: ^/trunk

Identification: JOSM/1.5 (12456 en) Linux Ubuntu 16.04.2 LTS
Memory Usage: 871 MB / 871 MB (395 MB allocated, but free)
Java version: 1.8.0_131-b11, Oracle Corporation, Java HotSpot(TM) 64-Bit Server VM
Screen: :0.0 1920x1080, :0.1 1366x768
Maximum Screen Size: 1920x1080
Dataset consistency test: No problems found

Last errors/warnings:
- W: java.net.SocketTimeoutException: Read timed out
- W: java.net.SocketTimeoutException: Read timed out
- W: java.net.SocketTimeoutException: Read timed out
- W: java.net.SocketTimeoutException: Read timed out
- W: No default layer selected, choosing first layer.
- W: No default layer selected, choosing first layer.
- W: java.net.SocketTimeoutException: Read timed out
- W: java.net.SocketTimeoutException: Read timed out
- W: Already here java.net.SocketTimeoutException: Read timed out
- W: java.net.SocketTimeoutException: Read timed out

[mkoniecz]

It may be worth checking whatever this warning makes any sense - tourism=attractiom, tourism=alpine_hut also should not raise warnings.

From looking at JOSM values for tourism tag I see not even single value where this warning makes sense.

[mkoniecz]

To explain: tourism route leading to alpine hut may involve exact mapping of path reaching alpine hut entrance. This is a valid tagging, also if alpine hut is mapped as node.

[mkoniecz]

@JOSM maintainers

Is it considered as a bug? It should be simple to make a patch but I prefer to check whatever it would be accepted.

[Don-vip]

It seems r10037 / r10071 does not work anymore.

[Klumbumbus]

Replying to Don-vip:

It seems r10037 / r10071 does not work anymore.

It works. It reduces the warning for guideposts to info level, which is fine in my eyes as maybe 95% of the guideposts are not in the middle of the way but beside the way. The attched example random_guidepost.jpg​ is a rather rare case.

For all other tourism nodes I agree that it can make sense to put it directly on the way if the node is the last waynode, i.e. the way ends there.

A tourism=hotel node in the middle of a highway=* should still get a validator message at warning level.

[mkoniecz]

which is fine in my eyes as maybe 95% of the guideposts are not in the middle of the way

It would mean that 5% (10k+*) of guideposts get false positive message. And in how many cases it would be an useful error?

• ​https://taginfo.openstreetmap.org/tags/information=guidepost

[Klumbumbus]

It is just info level. This means it is just an information, not an error.

I've already corrected a lot of guideposts, which are wrongly tagged on the way. (In some cases these are propably results of unwanted "node snapping" in the editors.) In non alpine areas there are usually no guideposts on the way at all.

[Don-vip]

I got it at warning level during my tests, not info level, I don't understand

[Klumbumbus]

Which tag combination?

[Don-vip]

OK understood. The code behaves as expected only of information level is enabled. Otherwise the rule is not applied, and the MapCSS class not set.

[Klumbumbus]

Oh, you're right. I didn't know that.

[Don-vip]

Me neither, I consider it a bug.

[Klumbumbus]

Well, we could simply split that rule into two, one without error output and one with error output, however we may fall into this trap again some day. Do you plan to fix this in the core?

[Don-vip]

Yes I'm working on a proper fix :)

[Don-vip]

In 12948/josm:

fix #15032 - MapCSS classes are not set if defined in information-level rules

[Klumbumbus]

I want to adjust the warning rules for end nodes.

[Klumbumbus]

In 13006/josm:

• fix #15032 - don't warn about tourism nodes on way end nodes
• see #15401 - adblue is not a fuel, restores sorted fuel list too