I had the desire for a new SVN command: svn sdiff. Does not exist, sadly, but svn is so CLI friendly at times I want to shake the hands of all the Subversion developers.
So:
svn cat some_file.c | sdiff -w220 - some_file.c
will do one file, and with mixed results you can do:
for i in `svn stat | egrep '^M' | awk '{ print $2 }'`; do svn cat $i | sdiff -w220 - $i; done
The downside is sdiff spits out whole files, so there's a lot of output; I long for grep's --before-context and --after-context. I don't demarcate the start of new files either, but a simple "echo $i;" should do it in the loop.