[mkgmap-dev] --index crashes on an otherwise successfully rendered image

Sat Nov 28 17:15:23 GMT 2009

Clinton Gladstone wrote:
> On Nov 27, 2009, at 11:09, Lambertus wrote:
>
>   
>> Mkgmap sometimes crashes with when trying to build a gmapsupp including 
>> index using an existing (prerenderd) image.
>>
>> Exception in thread "main" java.lang.IndexOutOfBoundsException: Index: 
>> 32803, Size: 488
>>         at java.util.ArrayList.rangeCheck(ArrayList.java:571)
>>         at java.util.ArrayList.get(ArrayList.java:349)
>>         at 
>> uk.me.parabola.imgfmt.app.lbl.LBLFileReader.readPoiInfo(LBLFileReader.java:354)
>>     
>
> I played around with this for a while, but could not find anything definite. I did find that the first POI which caused the problem was the following:
>
> P+R DOKTOR HOOLBOOMWEG
>
> (There were others which followed, which also caused this error.)
>
> It looked to me like the city value for the POI was incorrectly coded when the img file was originally generated.
>
> The best I can offer you as a temporary would be a few lines of code which would catch the error and not generate the city information. This would allow the map to compile, but the map would be potentially broken.
>
> Cheers.
Thanks for looking at the problem. I looked at the POI in Potlatch: 
seems perfectly alright to me.

If you catch the error can you print an warning/error? I don't 
understand why the map would potentially be broken? Is it possibile to 
leave the POI out entirely and would that prevent a broken map?

During my tests with a new transliterator for the Garmin maps I found 
out that there was a corruption of the XML source file. I'm not sure if 
this file was also used for generating that problem map, I'm almost sure 
it wasn't but I can't rule it out either.

Anyway, I'm running a complete new update now, perhaps the problem will 
be gone this time.