[mkgmap-dev] [PATCH v1] Decide in style file if a way is closed automatically

Fri Oct 26 17:52:46 BST 2012

wanmil wrote:
> I see: the message "Use both rules..." is printed out only if the way
> is
> not closed and both endpoints are outside the bounding box. So it is
> quite seldom. Anyhow I wonder why you don't see such a message. The
> message is generated before the style files are evaluated so your
> style
> rules have no influence on if the message is printed or not.

Ah ok, I have tested it first only with one big bbox for the whole Benelux so that makes sense.

I have now runned the patch on two tiles containing all osm data. It gives tons of line warnings "Use line rules only for WAY" for mkgmap:closed=false; They are way too many warnings to make sense of it, so I don't think those warnings are sensible at all so why print them out?

If I skip them there are 30 warnings left with "Use both rules"
Most of them are also not relevant (borders, highways) and just a few of them are waterways with mkgmap:closed=false & mkgmap:autoclosing=true which were indeed causing floodings before.

> Do you think the naming of the extra tags mkgmap:closed and
> mkgmap:autoclosing is good? Or do you have a better naming which is
> more
> intuitive? It's better to change that before committing :-)

I don't have a better suggestion for those tags. If the description is ok, I think it can be usefull (except for the warnings, but I assume you turn them off when committing it?)

Josef wrote:
> You should consider that many canal areas have the tag
> waterway=riverbank, or is this not relevant.
> This tag is false because there is stagnant water.

They are only handled as polygons which should be rendered ok as long they are closed.
This is not relevant in my case, where canal areas like this one was tagged only with waterway=canal:
http://www.openstreetmap.org/browse/way/115828742

I added natural=water to it. If I delete the tag waterway=canal, it will be in no matter of time changed by a Potlatch user who thinks that this is not a "lake" ;-) If you render it with mkgmap, this canal would be empty otherwise (unless the new patch will be comitted and the default water rules changed a bit).