For working with text that is prose, stored in plain text, I really like the output of git diff --color-words
, but I don't like the way it is dependent on ANSI terminal escape sequences. I'd like to get output that I can use to turn into HTML, some kind of user-friendly raw character display, or to generate statistics on the differences between the two files.
I prefer Perl (including CPAN), but would settle for pretty much anything.