Major Description
The scenario is that you have a bunch of Nutch data which has been gathered over some period of time. Some of the data structures are present, some are not. In segments directory for example there is .zip files (don't ask why) and in other directories there are .tar.gz files, etc.
This patch improves the SegmentChecker to skip directories or files present within the segments directory which are not 14 characters in length as ALL segments are. It also uses this check for individual segments if used by the IndexingJob. This means that we can prevent the Indexer blowing up if it is run on one segment (e.g. without -dir option) and detects some arbitrary directory present within segments/ which actually turns out not to be a segment afterall.