[mkgmap-dev] Branch is_in ready for a first test

Thu Jan 9 15:07:28 GMT 2020

Hi Nick,

yes and no. With r4008 the result is probably the same but performance can be worse compared to 'any' as the algorithm doesn't stop early.
The code contains e.g.
                if (statusFirst == Status.IN && "any".equals(mode))
                        return true;
which means "if the first point of the way is in we can return true" without checking any further.

I did not yet add checks regarding the correctness of the method parameter because they might change again. Will do this now.

Gerd

________________________________________
Von: Pinns UK <osm at pinns.co.uk>
Gesendet: Donnerstag, 9. Januar 2020 15:59
An: Gerd Petermann; mkgmap-dev at lists.mkgmap.org.uk
Betreff: Re: AW: AW: AW: AW: AW: [mkgmap-dev] Branch is_in ready for a first test

Hi Gerd

Yes I did

It must have defaulted to 'any'

Nick

On 09/01/2020 14:58, Gerd Petermann wrote:
> Hi Nick,
>
> did you really use some? The code expects either "any" or "all", but doesn't yet check this.
>
> Gerd
>
> ________________________________________
> Von: Pinns UK <osm at pinns.co.uk>
> Gesendet: Donnerstag, 9. Januar 2020 15:55
> An: Gerd Petermann; mkgmap-dev at lists.mkgmap.org.uk
> Betreff: Re: AW: AW: AW: AW: [mkgmap-dev] Branch is_in ready for a first test
>
> Hi Gerd
>
> I had to replace 'all' with 'some'  as it very cleverly picked out some
> small stretches of stream uniquely confined to a wood but it looked odd
> just to highlight these small waterways.
>
> 'some' doesn't do a bad job !
>
> Nick
>
> On 09/01/2020 14:30, Gerd Petermann wrote:
>> Hi Nick,
>>
>> river in wood polygon is probably a "good" worse case scenario for performance tests ;-)
>> Don't know how it is today, but a few years ago Japan was full of very complex wood MP (> 30.000 nodes) and long waterways...
>> I have no idea yet how that will perform...
>>
>> Gerd
>>
>> ________________________________________
>> Von: Pinns UK <osm at pinns.co.uk>
>> Gesendet: Donnerstag, 9. Januar 2020 15:24
>> An: Gerd Petermann; mkgmap-dev at lists.mkgmap.org.uk
>> Betreff: Re: AW: AW: AW: [mkgmap-dev] Branch is_in ready for a first test
>>
>> Hi Gerd
>>
>> Yes indeed, and also make streams etc  lighter, so it opens up a host of
>> new possibilities.
>>
>> Many thanks again for all your ( and Ticker's) hard work!
>>
>> Nick
>>
>> On 09/01/2020 14:21, Gerd Petermann wrote:
>>> Hi Nick,
>>>
>>> thanks for the quick feedback :)
>>>
>>> Interesting use case. Did not even think about this. Means you create overlay lines with different types for the routable way?
>>>
>>> Gerd
>>>
>>> ________________________________________
>>> Von: Pinns UK <osm at pinns.co.uk>
>>> Gesendet: Donnerstag, 9. Januar 2020 15:07
>>> An: Gerd Petermann; mkgmap-dev at lists.mkgmap.org.uk
>>> Betreff: Re: AW: AW: [mkgmap-dev] Branch is_in ready for a first test
>>>
>>> Hi Gerd
>>>
>>> It works a treat !
>>>
>>> Am using it with natural=wood/scrub which now enables me to make the
>>> colour of footpaths lighter and easier to see on a gps
>>>
>>> r
>>>
>>> Nick
>>>
>>> On 09/01/2020 13:57, Gerd Petermann wrote:
>>>> Hi Nick,
>>>>
>>>> yes, no extra option needed.
>>>>
>>>> Gerd
>>>>
>>>> ________________________________________
>>>> Von: Pinns UK <osm at pinns.co.uk>
>>>> Gesendet: Donnerstag, 9. Januar 2020 14:56
>>>> An: Gerd Petermann; mkgmap-dev at lists.mkgmap.org.uk
>>>> Betreff: Re: AW: [mkgmap-dev] Branch is_in ready for a first test
>>>>
>>>> Thanks Gerd
>>>>
>>>> That works
>>>>
>>>> So no extra --is-in  options required?
>>>>
>>>> r
>>>>
>>>> Nick
>>>>
>>>> On 09/01/2020 13:53, Gerd Petermann wrote:
>>>>> highway=* & bicycle!=* & is_in(landuse,cemetery,all)=true {add bicycle=dismount}