I originally downloaded version 0.7.13 and converted one odt file - or I might have converted it to doc first - to epub using Calibre. It gave me output consisting of 70 xhtml files, one for each chapter, which had names of the format partn.xhtml where n was a chapter number.
The next time I made a conversion I was using an updated version, and suddenly my 70 chapters had become 84 documents, with names of the format index_split_000.xhtml up to index_split_084.xhtml. Chapter 53 had suddenly become several chapters long.
It happens that this time I tested the debug function, and I can see under the Parsed folder that, whilst most paragraphs are marked <p class="P-P15">, the ones where the breaks appear are marked <p class="P-P7"> and <p class="P-P8">, both of which include auto page breaks. This chapter in the original document has not been changed between runs, and I can see no difference between the paragraphs in question and the surrounding ones. I've checked the area in Open Office, with View/Non-printing Characters, and there is just the usual paragraph mark at the end of each.
I have since tested 0.7.13 again, and this, too, now gives me the anomalous behaviour.
Any idea why this may be happening?
TIA