I do generally say that FFF doesn't change manually edited chapters, but that isn't entirely true.
There are a number of FFF's HTML processing steps that are applied to updated chapters, including image processing.
Image files included in FFF epubs are given normalized names, ffdl-0.jpg, ffdl-1.jpg, etc. We can't trust the original image names to be unique or reasonable length.
In order to recognize duplicate images, FFF records the image's original URL in the 'longdesc' attribute, eg: <img longdesc="origurl" ...> longdesc is a recognized, but little used attribute to make epub checkers happy.
Long story short, if an img tag is found on epub update without a longdesc attribute, FFF declares it a failed download and discards it. I'm reluctant to change that.
Searching the thread for longdesc,
this past post seems the most helpful.
Quote:
Originally Posted by JimmXinu
...
Put the <img> tags in the chapter text with src attr pointing to the URL of the image on disk (or where ever) and then run an FFF Update Always (if there aren't new chapters) and FFF will process it as if the author had put it in.
Your custom image will be 'downloaded', mapped to an ffdl-# name and included in the epub. |
In fact, it should be
easier for Elaine T's kid because they only need to change/copy the links to img tags and don't need to download and embedded the images.