[OSM-dev] using current planet with osm2pqsql

[Brett Henderson]

I don't know why some XML parsers are accepting this character and 
others aren't but I believe the correct behaviour is to throw an error.  
This is the same problem we had with the Java XML parser a couple of 
days ago in the thread "control characters in planet.osm".  o023 is 
equal to 0x13 and is an illegal XML character.  For now, it is simplest 
to strip it out with a sed command.

This character was originally entered by Potlatch.  A change has been 
committed to the Potlatch API to strip this character out (and others 
like it) to prevent it happening in the future.

Apologies if this is already known but it seems like we're wasting 
effort looking at xml parser behaviour when the real problem is invalid xml.

Jon Burgess wrote:
> On Mon, 2007-07-16 at 22:04 +0200, Joerg Ostertag wrote:
>   
>>>> I'd like to diagnose the error that you saw. Can you send me any
>>>> relevant error messages and library versions? The program works ok for
>>>> me.
>>>>         
>>> Worked fine for me too, I didn't notice any problems loading up last
>>> week's planet file, nor the week before.
>>>       
>> Maybe the XML-parser on our two machines are different?
>>     
>
> libxml2 is the only XML parser which is used by osm2pgsql unless you've
> made big changes to the source code. The error must be coming from
> libxml2. What version do you have installed?
>
> The 'tile.osm' machine is currently using 'libxml2-2.6.24.dfsg-1ubuntu1'
> which is a few minor versions behind the latest release.
>
> 	Jon
>
>
>
>
> _______________________________________________
> dev mailing list
> dev at openstreetmap.org
> http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/dev
>