Thx @dm101 for the files. I can see what the problem is (and indeed this is probably becky's issue as well - you mentioned pdf which is why I thought it may be different but it is an ePub you sent that showed the issue). It is when you have a file with those annoying embedded font-face declarations at the top like this:
<style type="text/css">
@font-face {
font-family: Courier;
panose-1: 2 7 4 9 2 2 5 2 4 4
}
I've never understood the point of these (and rip them out of my own ePubs). Obviously with enough of them in there the chances of hitting a number that coincidentally looks like an ISBN is higher.
I already have some code in there that rips out HTML tags. I will tweak that a bit to make sure these get ignored as well when evaluating.
|