Source code in hadoop-common repo has a bunch of files that have CRLF endings.
With more development happening on windows there is a higher chance of more CRLF files getting into the source tree.
I would like to avoid that by creating .gitattributes file which prevents sources from having CRLF entries in text files.
But before adding the .gitattributes file we need to normalize the existing tree, so that people when they sync after .giattributes change wont end up with a bunch of modified files in their workspace.
I am adding a couple of links here to give more primer on what exactly is the issue and how we are trying to fix it.
I will submit a separate bug and patch for .gitattributes