[DaltonST]

I just ran it against all of the books in one of my test libraries for Calibre 2.40 Windows 64bit, and it worked perfectly.

It looks to me that you are jumping to a conclusion without having posted any empirical data to prove that it is not the .epub's fault.

For example, you did not attach a screen-snip indicating that the "Count Pages" plug-in actually found real pages of "text". That is quite common in .PDF files that were created from scans, since "images" are not "text". "Count Pages" might find zero pages of "text" in a .PDF that is 5mb in size.

Your log showed no errors, and looks normal other than the fact that it extracted no text from the .epub. I suspect that your .epub has problems.

I suggest that you 'fix' your .epub by:

(1) reconverting it from an epub to an epub;

(2) running it against the excellent "Modify Epub" plug-in, clicking almost all of the checkboxes;

(3) running it against the "Count Pages" plug-in, confirming that is has "real" pages text, and is not just an .epub version of a scanned .PDF; and,

(4) converting the reconverted and "fixed" .epub to a .txt format, and then running ENF again for that book. ENF will use any .txt it finds before using a .epub format, and will use a .epub format before using a .PDF format. The log indicates that priority, and also will indicate which format it used.

After (4) above, open the .txt format in Notepad, and read it. Are there a large number of real English words? If there are, please PM me the .txt file so I can test with it.

Thanks.


DaltonST