[mkgmap-dev] r4836 stops Hungary & Romania

Tue Dec 28 13:22:33 GMT 2021

Hi Ticker,

okay, maybe you find time to implement a better solution.
I've committed the patch with r4838.

@Arndt: Please check if this solves the problems with your maps.
Not sure about this because you said mkgmap stops without any reaction
while this problem was a loop.

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, 28. Dezember 2021 13:20
An: Development list for mkgmap
Betreff: Re: [mkgmap-dev] r4836 stops Hungary & Romania

Hi Gerd

This fix is reasonable in the short term but has some problems:

If the invalid UTF16 string is output by other means (eg diagnostics),
it could cause a (more conventional) crash - but probably not looping.

Anything concatenated onto this string will be lost - this could be
fixed by initialising the encoder with
      onMalformedInput(CodingErrorAction.IGNORE);
instead of catching the error.

The full fix would required looking at all substr and character
indexing operations on string data from OSM tags and using
Character.is[High/Low/]Surrogate() or similar logic and adjusting the
start/end char index or whatever

Ticker

On Tue, 2021-12-28 at 07:23 +0000, Gerd Petermann wrote:
> Hi all,
>
> I was able to reproduce with option --latin1 and the modified default
> style.
>
> The offending node seems to be
> https://www.openstreetmap.org/node/9122388694
>
> mkgmap loops in method AnyCharsetEncoder.encodeText(String text)
> I guess the same can happen when labels are cut at length 170.
>
> The attached patch seems to fix this, but I have no idea if this is
> the right approach.
>
> Gerd
>
> ________________________________________
> Von: mkgmap-dev <mkgmap-dev-bounces at lists.mkgmap.org.uk> im Auftrag
> von Ticker Berkin <rwb-mkgmap at jagit.co.uk>
> Gesendet: Montag, 27. Dezember 2021 20:06
> An: Development list for mkgmap
> Betreff: Re: [mkgmap-dev] r4836 stops Hungary & Romania
>
> Looks like uft16 surrogate pair chars are being separated by the
> substr.
>
> Ticker
>
>
>
> On Mon, 2021-12-27 at 19:28 +0100, Arndt Röhrig wrote:
> >  ...mayby is in the osm data a kryptic text like this:
> >
> >  https://www.openstreetmap.org/node/9115233473
> >
>
> _______________________________________________
> mkgmap-dev mailing list
> mkgmap-dev at lists.mkgmap.org.uk
> https://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
> _______________________________________________
> mkgmap-dev mailing list
> mkgmap-dev at lists.mkgmap.org.uk
> https://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev

_______________________________________________
mkgmap-dev mailing list
mkgmap-dev at lists.mkgmap.org.uk
https://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev