[mkgmap-dev] Basecamp and NET/NOD changes

Wed Nov 27 12:17:36 GMT 2019

Hi Gerd

With a map generated without routing-island-removal, there are
geographic areas where Basecamp routing fails, but I couldn't find any
routing problems in the same areas with my eTrex or GPSMapEdit.

I can't to any more today or tomorrow, but later I'll have a go at
using the default style and reducing the area.

My style adds lots of circular walkways that are often not connected to
the main road network but, looking at the code, StyledConverted goes to
some trouble to split these in 2 and to start with a routing node.

What other cases are there where the routing node won't be the first on
the road, which is what seems to be expected by Display and, maybe,
Basecamp. one-way to/from nowhere?

Ticker

On Tue, 2019-11-26 at 16:58 +0000, Gerd Petermann wrote:
> Hi Ticker,
> 
> the reason for way -1477104 is simple:
> It is the only island when you use the default style and it appears
> as first way in the file. With option --preserve-element-order
> this seems to be a worst case scenario for BaseCamp. My understanding
> is that BaseCamp builds internal tables for road data
> and NOD data and that this fails when a road without NOD is processed
> before one that has NOD.
> So, my goal was to sort the data so that BC reads all the roads with
> NOD first. I gave up because I have no idea how exactly BC reads the
> data. It seems to ignore the order in NET, it seems to read the RGN
> file and uses the pointers to NET in it.
> 
> Gerd
> 
> ________________________________________
> Von: mkgmap-dev <mkgmap-dev-bounces at lists.mkgmap.org.uk> im Auftrag
> von Ticker Berkin <rwb-mkgmap at jagit.co.uk>
> Gesendet: Dienstag, 26. November 2019 17:31
> An: Development list for mkgmap
> Betreff: Re: [mkgmap-dev] Basecamp and NET/NOD changes
> 
> Hi Gerd
> 
> I've trying various things, like changing the TYP, removing road
> -speed/class, ... to try and pin down this problem and generally, the
> behaviour got worse.
> 
> I experimented a bit with your cut-down area: ticker-part-mod.zip but
> had problems. Way -1477104 confused me for a while. So I went back to
> the full tile.
> 
> I've just updated to trunk latest r4387, removed all my changes and
> did
> a clean rebuild, then, with --check-routing-island=-1 tried lots of
> Motorcar routing in Basecamp and I find areas of my map still
> misbehave.
> 
> EG:
> "so23 0je" > "so23 0ps" straight line to a road a few roads east then
> straight line back.
> "so23 0qb" > "so23 9ru" straight line.
> 
> Generally, all routing fails around this small area, but other areas
> north and west work fine.
> 
> The setup and data is identical to before but I've just been using
> the
> northern tile (74210002.osm.pbf)
> 
> I'm just going to try some of these on the Device and pick some other
> areas with island=-1 on the ticker-part-mod
> 
> Ticker
> 
> 
> On Mon, 2019-11-25 at 07:16 +0000, Gerd Petermann wrote:
> > Hi Ticker,
> > 
> > okay, I was not able to reproduce the problem on my Oregon, and it
> > turned out that the one routing problem in Mapsource was my fault,
> > I
> > asked for a route which was forbidden because of oneway roads.
> > I think that means that nobody should use option --check-routing
> > -island-len=INTEGER when they use BaseCamp or when they publish
> > their
> > maps unless we find a better implementation.
> > 
> > Gerd
> 
> _______________________________________________
> mkgmap-dev mailing list
> mkgmap-dev at lists.mkgmap.org.uk
> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev