If anybody is using Lucene 4.2 and wants a version of Luke that supports it I've got a version that supports 4.2 on my Github ( https://github.com/mpoindexter/luke). This tarball should have all the source and binaries:
https://github.com/mpoindexter/luke/archive/4.2-1.tar.gz I've also enhanced the "Reconstruct document" functionality a little bit: if there is no stored value for a field it now tries a little harder to find data to reconstruct. In addition to the current support for extracting terms for fields indexed with positions, it will now also try to read from doc values, process fields that were indexed without positions (all terms end up at position 0), and try to guess if a field was indexed as numeric or not, and extract the number if it thinks it was. If you want a version without these changes you should be able to pretty easily fork it on Github and apply just the commits relating to 4.2 support.
