a bit more grouping in the validator tree

[Klumbumbus]

Some validator tests print the affected tag and could therefore be grouped to clean up the validator tree. To avoid too much groups (too much clicking to finaly see the affected object), this should be done only for common mapping errors. I suggest groups for:

• deprecated tags ("{0} is deprecated")
• unconnected nodes ("{0} must be connected to a way")
• missing tags ("{0} without {1}")
• suspicious tag combination ("{0} together with {1}")

[Klumbumbus]

In 11138/josm:

fix #13804 - group deprecated tags and isolated nodes in the validator tree

[Klumbumbus]

In 11139/josm:

see #13804 - group deprecated tags

[Klumbumbus]

In 11146/josm:

see #13804 - a bit more grouping in the validator tree, small adjustments to a few validator rules

[Klumbumbus]

In 11153/josm:

• add display_values and values_sort=false for Route network values (iwn,nwn,...)
• warn about missing sport key on pitches
• see #13804 - group unnecessary tag warnings
• add own colors for sport=multi|running|athletics to avoid false positive validator warning in stadiums (inner style equals MP style)
• fix #13824 - adjust sport=equestrian|horse_racing presets

[Klumbumbus]

In 11353/josm:

see #13804 - move parking in "{0} inside {1}" validator group

[skyper]

• How about grouping the role checks according to the type=* ?
• Like to have the informal fixme/FIXME=*grouped by the value, that all same values are in a subgroup.

[Klumbumbus]

Replying to skyper:

• How about grouping the role checks according to the type=* ?

I think we should not create too much subgroups. The user already has to expand 3 folders until he finally sees the warning message (priority->group->subgroup->item).

• Like to have the informal fixme/FIXME=*grouped by the value, that all same values are in a subgroup.

I think that would create too much subgroups with most of them containing only one item.

[skyper]

Replying to Klumbumbus:

Replying to skyper:

• How about grouping the role checks according to the type=* ?

I think we should not create too much subgroups. The user already has to expand 3 folders until he finally sees the warning message (priority->group->subgroup->item).

• Like to have the informal fixme/FIXME=*grouped by the value, that all same values are in a subgroup.

I think that would create too much subgroups with most of them containing only one item.

My problem is that I will not look at 100s of single warnings. I need to filter them. E.g. I am willing to fix warnings about associatedStreet relations but ATM I always have to skip all the false positives of public-transport relations. Same is true for fixme=* where I won't fix them as I have to select every object on its own to get the value.

By the way, the warning descriptions are quite long anyway, that you always have to unpin the validator toogle dialog in order to work with it, so one more level does not matter in my point of view.

How about only subgrouping from a certain point of hits like five or ten warnings and keeping it flat for less ?

[Klumbumbus]

Replying to skyper:

My problem is that I will not look at 100s of single warnings. I need to filter them. E.g. I am willing to fix warnings about associatedStreet relations but ATM I always have to skip all the false positives of public-transport relations. Same is true for fixme=* where I won't fix them as I have to select every object on its own to get the value.

By the way, the warning descriptions are quite long anyway, that you always have to unpin the validator toogle dialog in order to work with it, so one more level does not matter in my point of view.

I did not mean the lenght of the texts. My concern is that it is annoying to open too much folders until you finally get your item.

I never unpin this dialog. The important information is usualy at the beginning of the text. (however sometimes I use the horizontal scrollbar).

The items in the role check are already grouped, usually there is a low number of groups and it is easy to spot which belong to routes or other relations.

Regarding the false positives, we should better fix them if you name them.

Regarding fixme, maybe it would be useful to directly display the fixme value in the warning text (after #14082 is done).

Anyway, all the suggestions should be better discussed in new tickets. This is an old one.

[Klumbumbus]

In 14987/josm:

• see #2760 see #13804 - fix validator rule syntax
• see #17592 - fix validator rule syntax
• fix #17594 - warn about tracktype=grade2 with surface=sand|mud