Marco Lechner - FOSSGIS e.V. wrote:
> An interesting additional behavior:
>
> If I use
> bzcat ../planet-090421.osm.bz2 | ./bin/osmosis --read-xml-0.6 file=-
> --bounding-polygon-0.6 file=../aoi.pff --write-xml-0.6
> file=../aoi_2009-04-21_v06.osm
> instead of
> ./bin/osmosis --read-xml-0.6 file=path/planet-090421.osm.bz2
> compressionMethod=bzip2 --bounding-polygon-0.6 file=path/aoi.pff
> --write-xml-0.6 file=path/aoi_2009-04-21_v06.osm
> it seems to work (still running).
> Although using bzcat has a better performance, the build-in function
> should work, shouldn't it? But there seems to be a bug. Is this a bug?
> Can anybody reproduce this?
>   
Marco, I've been able to reproduce the bug but I don't know how to fix 
it.  The bzip2 application used to compress the planet file uses 
parallel processing across multiple cores and I have a sneaking 
suspicion that it processes chunks of the file independently then just 
appends them together.  As a result the java bzip2 implementation that 
I'm using hits what it thinks is the end of the compressed stream (when 
it reality it is only the end of the first chunk) and stops processing.  
There may be a way of coding around this but it would take me a long 
time to figure out and it's not on top of the priority list right now.  
I'll gladly welcome a patch if somebody finds a solution.

I can only suggest that you continue to use the bzcat command.

Brett


_______________________________________________
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev

Reply via email to