Quote:
Originally Posted by Manichean
I'm guessing that the XPath works in the second ePub->MOBI step? In that case, the original ePub would have some deviation from what Calibre expects that breaks the XPath. The first ePub->ePub conversion would fix that deviation, so it works in the ePub->MOBI step. Try taking a close look at the structure of your original ePubs and comparing that with what the XPath expects.
|
no the xpath seems to work in the 1st step IF output type is EPUB. after step 1 ( inspecting in sigil )- the epub structure is one file per chapter.
I can then turn xpath off in the 2nd conversion as the book is already sorted.
I still don't get the difference between 1 pass epub to epub _ all OK, and 1 pass epub to mobi - not OK, starting both times with the exact same source.
as I understand calibre pipeline, the xpath stuff is in both cases done on intermediate HTML, before handing off to the output module ?