[OSM-talk] Missing day replicate

mmd mmd.osm at gmail.com
Sat Feb 9 11:47:26 UTC 2019


Am 09.02.19 um 08:17 schrieb Jochen Topf:
> 
> this might or might not be related to the hourly change file with the
> number 56107 being "strange": Osmosis can not read the file 107.osc.gz,
> it reports a UTF-8 error. But if I gunzip the file, Osmosis can read the
> file fine. So this points to an error in Osmosis' handling gzip'ed
> files. Unfortunately Osmosis is no longer maintained.
> 

For some strange reason, the gz file claims to be from "FAT filesystem",
and most likely osmosis can't figure out that the file contents are in
UTF-8.

Once I decompress and re-compress the file on Linux, osmosis seems to be
just fine with said file:


$ wget https://planet.openstreetmap.org/replication/hour/000/056/107.osc.gz

$ file 107.osc.gz
107.osc.gz: gzip compressed data, from FAT filesystem (MS-DOS, OS/2, NT)

$ ./osmosis --read-xml-change file=107.osc.gz --write-xml-change
file=107.osc.new
...
Caused by: org.xml.sax.SAXParseException; lineNumber: 243521;
columnNumber: 1; Invalid byte 2 of 4-byte UTF-8 sequence.
...

$ gunzip 107.osc.gz

$ gzip 107.osc

$ file 107.osc.gz
107.osc.gz: gzip compressed data, was "107.osc", last modified: Wed Feb
 6 02:02:08 2019, from Unix

$ ./osmosis --read-xml-change file=107.osc.gz --write-xml-change
file=107.osc.new
-> works fine

-- 




More information about the talk mailing list