[mkgmap-dev] [PATCH v2] Fix so that all polygons are really closed

Sun Feb 27 20:59:26 GMT 2011

On Sun, Feb 27, 2011 at 12:59:09PM +0100, WanMil wrote:
>> If I disable this rule altogether, the warning count drops from 2386 
>> to 476. That is, the selective suppression of the warnings would help 
>> a lot.
>
>Mmh, I am not a big fan of suppressing warning messages. But I see that 
>you have the problem that you only want to use man_made=pier with 
>polygons. If you add a "dead" rule to the line style you will loose all 
>man_made=pier. Would it help you if I add the tags of the way to the 
>warning message? Then you could easily filter the message.

I see that you implemented the tag list printout. It might be good 
enough.

>> One more idea: you might add some extra treatment for polygons that 
>> are more than 1 "lap". For example, 
>> http://www.openstreetmap.org/browse/way/76559964 comprised of nodes 
>> 902322231, 902322232, ..., 902322231, 902322232. The second 902322232 
>> triggered the error.
>
>Is this really a big problem? It is an error in the OSM data and just 
>in case this does not happen very often we should not support such 
>errors by fixing them automatically.

It is probably not a big problem, and I was not thinking that we should 
try to fix any garbage automatically. It could be nice to have the 
message distinguish incorrectly closed polygons from unclosed polygons.

It took me some time to figure out the problem. I was expecting to see a 
gap between some points, but there was none. Only after I split the way 
in JOSM, the highlighting of the area hinted me where the starting point 
was.

	Marko