Version mkgmap-r3876 was committed by gerd on Sun, 02 Apr 2017
further decrease peak memory and improve speed for global index creation
Correct the calculation of needed bytes for sort keys so that padding bytes are
considered. Now copies are only created when the help to reduce memory.
Thanks, i will fix the overlapping tomorrow and test again...
thomas
Am 01.04.2017 um 17:02 schrieb Gerd Petermann:
Hi Thomas,
I've changed splitter to print better information about the overlapping areas,
please try r583,
it is now also able to parse the file
Version mkgmap-r3875 was committed by gerd on Sat, 01 Apr 2017
decrease peak memory and improve speed for global index creation
- use memory cache were it helps and free memory for caches after SortKey are
calculated
- always free memory for mdr19 data
- implement (undocumented) --x-mdr7-excl
Hi, I have maked a new split with splitter r-440. And splitter r-440
worked without error. Espacially all nodes of multipolygons have the
tag=value like as expected.
thomas
Am 01.04.2017 um 13:39 schrieb Gerd Petermann:
Hi Thomas,
splitter r580 cannot read this file without a few small
Hi Thomas,
I've changed splitter to print better information about the overlapping areas,
please try r583,
it is now also able to parse the file http://img2ms.de/bilder/areasEuropa.list
With r583 I see these messages for that file:
...
Waring: The areas given in --split-file are overlapping.
Hi Thomas,
splitter r580 cannot read this file without a few small changes (blanks
added/removed).
After these changes it complains that there are overlaps.
Seems you are using a different file or a different program.
Gerd
Von: mkgmap-dev
I have controled my areasEuropa.list (download
:img2ms.de/bilder/areasEuropa.list). I found no overlapping around the
5707. All neighbours 5705; 5709; 5728; 5729;
5708 ;5706; 5705) have identical borders without any
overlapping. Or shout it have a overlapping ?
Hi Thomas,
just make sure to keep the splitfile it that really caused the problem with
r580.
In the moment I prefer to think that you used r429 by mistake ;-)
Gerd
Von: mkgmap-dev im Auftrag von Thomas
OK - again, I misunderstood something.
Ticker
On Sat, 2017-04-01 at 10:09 +, Gerd Petermann wrote:
> Hi Ticker,
>
> I don't think that my change is much different to yours.
> The current code is
> if (lat != lastLat || lon != lastLon
> ||
>
Hi Thomas,
the problem was caused by your manually changed split file which contained
small overlaps. Do you still use that one?
I think I fixed the problems with that one in r440.
Or maybe you created a new one which causes new problems?
Gerd
Von:
Hi Ticker,
I don't think that my change is much different to yours.
The current code is
if (lat != lastLat || lon != lastLon ||
(checkPreserved && (p instanceof CoordNode
|| p.preserved(
I was not sure if we have a case
Hi Gerd, the old problem exists till now. I splitt europa-latest-osm.pbf
using splitter r-580. But splitterr-580 do not proper splitt the
multipolygon /("Schloßteich", 3 Elemente) [Kennung: 59.766]/ . The mp
has tags: /name=Schloßteich; natural=water, type=multipolygon/. This mp
has
Hi Gerd
I think your fix will remove most of the benefits of the
PredictFilterPoints, by keeping all original map coords.
My mistake was assuming that only original points were candidates for
preserving but this looks not to be true. It looks like housenumber
processing generates points to be
Thansk for the quick fix
El 01/04/17 a las 07:39, Gerd Petermann escribió:
Hi Carlos,
this problem is fixed with r3874. It was introduced with the merge from the
split-shape branch.
Not sure why it only occurs with this special style.
Gerd
Von:
14 matches
Mail list logo