I worked through the RTF generated by Word, and this does cause ebook-convert to generate a page break if it's converting to .mobi, but strangely I still get no page break when converting to .pdf:
Code:
{\rtf1\ansi
{\fonttbl{\f0\fswiss\fprq2\fcharset0 Arial;}}
{\colortbl ;\red0\green0\blue0;}
\viewkind4\uc1\pard\sa200\cf1\lang2057\f0
\fs28 Citizen of the Galaxy\fs24
\par Robert A. Heinlein
\par Copyright 1957
{\par \page }
\fs28 Chapter 1\fs24
\par "Lot ninety-seven," the auctioneer announced. "A boy."
}
The only difference is using:
{\par \page }
\fs28 Chapter 1\fs24
instead of:
\page\par\fs28 Chapter 1\fs24
though why that should make a difference I don't know.
I read up on XPath, but as far as I can see it's a method for identifying nodes in XML. If I wanted to use --page-breaks-before would I have to convert to XML then convert to PDF?