generate-sea:polygons broken between mkgmap r2160 and r2168

classic Classic list List threaded Threaded
11 messages Options
Reply | Threaded
Open this post in threaded view
|

generate-sea:polygons broken between mkgmap r2160 and r2168

Charlie Ferrero
Hi list,

Just upgrading mkgmap to keep up to date and I find that
--generate-sea:polygons has been broken somewhere between mkgmap-r2160
(which works) and mkgmap-r2168 (which doesn't).

Screenshot 1: created with mkgmap-r2160
http://cferrero.net/maps/img/r2160.png
Screenshot 2: created with mkgmap-r2168
http://cferrero.net/maps/img/r2168.png

As you can see, by r2168 the sea is no longer being drawn (the blue bits
that remain are actually mangrove or wetland polygons).

I used the same command for screenshots 1 & 2:
--generate-sea:polygons,extend-sea-sectors,land-tag=natural=background

Can someone see what bug was introduced?

--
Charlie
_______________________________________________
mkgmap-dev mailing list
[hidden email]
http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Reply | Threaded
Open this post in threaded view
|

Re: generate-sea:polygons broken between mkgmap r2160 and r2168

WanMil
Hi Charlie,

I have no idea yet.

I think there are only two commits that may be relevant:
r2163 and r2168.
It would be great if you can retry with r2163 to narrow down the
problem. You might also upload the OSM data of your tile so that I or
someone else can do some debugging to find the problem.

WanMil

> Hi list,
>
> Just upgrading mkgmap to keep up to date and I find that
> --generate-sea:polygons has been broken somewhere between mkgmap-r2160
> (which works) and mkgmap-r2168 (which doesn't).
>
> Screenshot 1: created with mkgmap-r2160
> http://cferrero.net/maps/img/r2160.png
> Screenshot 2: created with mkgmap-r2168
> http://cferrero.net/maps/img/r2168.png
>
> As you can see, by r2168 the sea is no longer being drawn (the blue bits
> that remain are actually mangrove or wetland polygons).
>
> I used the same command for screenshots 1&  2:
> --generate-sea:polygons,extend-sea-sectors,land-tag=natural=background
>
> Can someone see what bug was introduced?
>
> --
> Charlie
> _______________________________________________
> mkgmap-dev mailing list
> [hidden email]
> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev

_______________________________________________
mkgmap-dev mailing list
[hidden email]
http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Reply | Threaded
Open this post in threaded view
|

Re: generate-sea:polygons broken between mkgmap r2160 and r2168

Charlie Ferrero
On 18/02/2012 01:02, WanMil wrote:

> Hi Charlie,
>
> I have no idea yet.
>
> I think there are only two commits that may be relevant:
> r2163 and r2168.
> It would be great if you can retry with r2163 to narrow down the
> problem. You might also upload the OSM data of your tile so that I or
> someone else can do some debugging to find the problem.
>
> WanMil
>
Can you send a link to r2163? http://www.mkgmap.org.uk/snapshots/ only
goes back to r2168

The original OSM files are here:
http://cferrero.net/maps/downloads/63251501.osm.pbf
http://cferrero.net/maps/downloads/63251502.osm.pbf


--
Charlie

_______________________________________________
mkgmap-dev mailing list
[hidden email]
http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Reply | Threaded
Open this post in threaded view
|

Re: generate-sea:polygons broken between mkgmap r2160 and r2168

Sterol Andro
Hello, Charlie.
I noticed the same issue, but I thought it was my fault... I am not very skilled in TYPs and styles..

So I compiled the mkgmap versions between 2160 and 2169 for your perusal.
You can find them here:


Just let me know if I can help in any way...




On Sat, Feb 18, 2012 at 9:22 AM, Charlie Ferrero <[hidden email]> wrote:
On 18/02/2012 01:02, WanMil wrote:
> Hi Charlie,
>
> I have no idea yet.
>
> I think there are only two commits that may be relevant:
> r2163 and r2168.
> It would be great if you can retry with r2163 to narrow down the
> problem. You might also upload the OSM data of your tile so that I or
> someone else can do some debugging to find the problem.
>
> WanMil
>
Can you send a link to r2163? http://www.mkgmap.org.uk/snapshots/ only
goes back to r2168

The original OSM files are here:
http://cferrero.net/maps/downloads/63251501.osm.pbf
http://cferrero.net/maps/downloads/63251502.osm.pbf


--
Charlie

_______________________________________________
mkgmap-dev mailing list
[hidden email]
http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev


_______________________________________________
mkgmap-dev mailing list
[hidden email]
http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Reply | Threaded
Open this post in threaded view
|

Re: generate-sea:polygons broken between mkgmap r2160 and r2168

Charlie Ferrero
In reply to this post by WanMil
On 18/02/2012 01:02, WanMil wrote:

> Hi Charlie,
>
> I have no idea yet.
>
> I think there are only two commits that may be relevant:
> r2163 and r2168.
> It would be great if you can retry with r2163 to narrow down the
> problem. You might also upload the OSM data of your tile so that I or
> someone else can do some debugging to find the problem.
>
> WanMil
> tp://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Hey WanMil,

The problem was introduced by mkgmap r2163 (thanks to Sterol Andro for
the individual jar files).  r2162 compiles fine with sea visible; r2163
shows the no sea problem described earlier when the
--generate-sea:polygons option is used.

I guess the preprocessing introduced by r2163 is somehow removing
something that the --generate-sea:polygons algorithm depends on.

--
Charlie

_______________________________________________
mkgmap-dev mailing list
[hidden email]
http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Reply | Threaded
Open this post in threaded view
|

Re: generate-sea:polygons broken between mkgmap r2160 and r2168

WanMil
Thanks all for tracking that down.
I can confirm that.

The problem is the new UnusedElementsRemoverHook which removes all
elements that are outside the tile. Unfortunately it also removes the
sea polygon which is one point bigger than the bounding box. I will
think about how to fix that but it should not be too complicated.

WanMil

> On 18/02/2012 01:02, WanMil wrote:
>> Hi Charlie,
>>
>> I have no idea yet.
>>
>> I think there are only two commits that may be relevant:
>> r2163 and r2168.
>> It would be great if you can retry with r2163 to narrow down the
>> problem. You might also upload the OSM data of your tile so that I or
>> someone else can do some debugging to find the problem.
>>
>> WanMil
>> tp://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
> Hey WanMil,
>
> The problem was introduced by mkgmap r2163 (thanks to Sterol Andro for
> the individual jar files).  r2162 compiles fine with sea visible; r2163
> shows the no sea problem described earlier when the
> --generate-sea:polygons option is used.
>
> I guess the preprocessing introduced by r2163 is somehow removing
> something that the --generate-sea:polygons algorithm depends on.
>

_______________________________________________
mkgmap-dev mailing list
[hidden email]
http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Reply | Threaded
Open this post in threaded view
|

Re: generate-sea:polygons broken between mkgmap r2160 and r2168

toc-rox
This post was updated on .
In reply to this post by Charlie Ferrero
This works for me:

generate-sea:multipolygon,no-sea-sectors,extend-sea-sectors,close-gaps=5000,land-tag=natural=land

Klaus

Edit: Sorry, you use "polygon" and not "multipolygon".
Reply | Threaded
Open this post in threaded view
|

Re: generate-sea:polygons broken between mkgmap r2160 and r2168

Charlie Ferrero
On 18/02/2012 18:14, toc-rox wrote:
> This works for me:
>
> generate-sea:multipolygon,no-sea-sectors,extend-sea-sectors,close-gaps=5000,land-tag=natural=land
>
> Klaus
>
> --
>
Yes, multipolygon works fine, but I need the polygon version.

--
Charlie

_______________________________________________
mkgmap-dev mailing list
[hidden email]
http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Reply | Threaded
Open this post in threaded view
|

Re: generate-sea:polygons broken between mkgmap r2160 and r2168

WanMil
In reply to this post by WanMil
I've comitted a fix to the UnusedElementsRemoverHook so it should work
again.

Thanks for the good report!!
WanMil

> Thanks all for tracking that down.
> I can confirm that.
>
> The problem is the new UnusedElementsRemoverHook which removes all
> elements that are outside the tile. Unfortunately it also removes the
> sea polygon which is one point bigger than the bounding box. I will
> think about how to fix that but it should not be too complicated.
>
> WanMil
>
>> On 18/02/2012 01:02, WanMil wrote:
>>> Hi Charlie,
>>>
>>> I have no idea yet.
>>>
>>> I think there are only two commits that may be relevant:
>>> r2163 and r2168.
>>> It would be great if you can retry with r2163 to narrow down the
>>> problem. You might also upload the OSM data of your tile so that I or
>>> someone else can do some debugging to find the problem.
>>>
>>> WanMil
>>> tp://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
>> Hey WanMil,
>>
>> The problem was introduced by mkgmap r2163 (thanks to Sterol Andro for
>> the individual jar files).  r2162 compiles fine with sea visible; r2163
>> shows the no sea problem described earlier when the
>> --generate-sea:polygons option is used.
>>
>> I guess the preprocessing introduced by r2163 is somehow removing
>> something that the --generate-sea:polygons algorithm depends on.
>>
>
> _______________________________________________
> mkgmap-dev mailing list
> [hidden email]
> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev

_______________________________________________
mkgmap-dev mailing list
[hidden email]
http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev

noremove_large_ways_v1.patch (2K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: generate-sea:polygons broken between mkgmap r2160 and r2168

toc-rox
In reply to this post by Charlie Ferrero
I have experimented with "polygons" in the past but this was unsuccessful.
May be due to the problem you have found yet.

Could you explain the basic difference between "multipolygon" and "polygons" ?
And what is the advantage of "polygons" ?

Thanks - Klaus
Reply | Threaded
Open this post in threaded view
|

Re: generate-sea:polygons broken between mkgmap r2160 and r2168

Charlie Ferrero
toc-rox ([hidden email]) wrote:

> I have experimented with "polygons" in the past but this was unsuccessful.
> May be due to the problem you have found yet.
>
> Could you explain the basic difference between "multipolygon" and "polygons"
> ?
> And what is the advantage of "polygons" ?
>
> Thanks - Klaus
>

The advantage is quite specific, but for older GPS units where you  
cannot change the background polygon colour,  
--generate-sea:multipolygon creates maps which display land as "Garmin  
yellow".  --generate-sea:polygons on the other hand creates a land  
polygon which you can specify in a TYP file, which is much nicer. :)

--
Charlie

_______________________________________________
mkgmap-dev mailing list
[hidden email]
http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev