"Possible answer invalidation" has too many false positives

[retailcoder]

Duga seems to link to every single edit to any answered question.

She needs to be pickier than that if she wants to claim possible answer invalidation - otherwise the message could be a notch less alarming if changed to post-answer edit.

But I like the idea of picking up possible answer invalidation edits. The tool just needs a little fine-tuning, to ignore edits where no code is involved.

[Zomis]

There are two problems here that I can see:

• Simple backtick code stuff is included in the check for whether or not code was changed, see revision 6 at http://codereview.stackexchange.com/posts/17758/revisions
• Intendation changes, see revision 10 at http://codereview.stackexchange.com/posts/3668/revisions

[Zomis]

https://github.com/Zomis/Duga/commit/09bfb4bdcdf8c3cdb240a75a9f3383d314b87761 and other recent commits should have fixed this