[Tex2002ans]

Originally Posted by jackie_w

Perhaps you didn't look hard enough. Every <span> with class="CharOverride-4" is italic. There are hundreds of those.

Hence the more surgical approach—deleting known useless style info (absolute positioning), then Calibre EPUB->EPUB conversion.

It's how I typically deal with really atrocious InDesign->EPUBs or Google Docs->EPUB. (If you think InDesign is bad, you should see Google Docs's code!)

Note: Like a few InDesign oddities I've been running across lately:

In some files, InDesign assigns an "Italic" font, but uses "font-style: normal" within the CSS.

So in a clean document, it should look like this:

<p>This is an <i>italic</i> word.</p>

but InDesign->EPUB may appear like this:

<p>This is an <span class="CharOverride-4">italic</span> word.</p>

span.CharOverride-4 {
	font-family:"Minion Pro Italic";
	font-style:normal;     <----- See here. Should say italic.
	font-weight:normal;
}

I'm unsure of how/why...

I've also run across similar in some DOCXs I've gotten from authors. Toxaris's EPUBTools doesn't handle these types of wrongly-styled fonts properly, but he is aware of the issue (and I believe he was working on it for the next release).

If you just mass throw away <span>s/fonts without looking closely, you'd also lose this sort of information.