Anyone else have a workaround for reusing an input stream that has been given to Tika Detect?
According to inline comments in the Tika code, it gives the impression that developers understand about how to correctly use an Input Stream, but the implementation doesn't appear to match the comments. i.e. use mark, reset. I've written a sample test, https://gist.github.com/nhojpatrick/ac01a3b3d791364b26f8 Compile it then do java -cp ./ TikaDetectTester [path to file to test]. It will create a *.out1 of the origional, then do a detect then write the file out again to *.out2 using the input stream given to tika. Cheers, John
