Convert diff to markdown with strikeout?

algorithm, diff, markdown

Solution

Use wdiff. It already does the word-by-word comparison you're looking for; converting its output to markdown should take just a few simple regular expressions.

For example:

$ cat foo
Why do we study programming languages?  Not in order to
$ cat bar
We study programming languages not in order to
$ wdiff foo bar
[-Why do we-]{+We+} study programming [-languages?  Not-] {+languages not+} in order to
$ wdiff foo bar | sed 's|\[-|<em>|g;s|-]|</em>|g;s|{+|<strike>|g;s|+}|</strike>|g'
<em>Why do we</em><strike>We</strike> study programming <em>languages?  Not</em> <strike>languages not</strike> in order to

Edit: Actually, wdiff has some options that make it even easier:

$ wdiff -w '<em>' -x '</em>' -y '<strike>' -z '</strike>' foo bar
<em>Why do we</em><strike>We</strike> study programming <em>languages?  Not</em> <strike>languages not</strike> in order to

Problem

I'd like to convert the output of `diff` (on a Markdown file) to Markdown with `<strike>` and `<em>` tags, so that I can see what has been removed from or added to a new version of a document. (This kind of treatment is very common for legal documents.) Example of hoped-for output: Why do weWe study programming languages? notNot in order to ... One of the many difficulties is that diff's output is line-oriented, where I want to see differences in individual words. Does anyone have suggestions as to what algorithm to use, or what software to build on?

Original source