new mapreduce API and NLineInputFormat

[Henning Blohm]

I am using the new mapreduce.* API in my jobs on hadoop 0.20.2. - I
actually 
have some utilities around job scheduling and such.

Now I would like to use NLineInputFormat to parallelize some data import
task into a number of mappers.

However in 0.20.2 you cannot call

job.setInputFormatClass(NLineInputFormat.class);

as NLineInputFormat does not extend the "right" InputFormat interface
(in 
contrast to the 0.21 version).

Is there any way to make this work on 0.20.2? 

Thanks,
  Henning