org.apache.hadoop.mapreduce.Partitioner access to JobConfiguration

[Julien Le Dem]

I'm looking at hadoop 0.18 vs 0.20.
Why doesn't Partitioner have a setup(JobContext) method ?
It seem before it used to be a JobConfigurable.
Parameters can be passed statically by the Mapper but it's not ideal.
Someone can enlighten me ?
Julien