Vexing error in EPUB file created in Sigil

[pstjmack]

I'm struggling to fix the few outstanding glitches in the EPUB version of my latest book, created in Sigil. There's only one remaining error warning, and it reads like this:

"WARNING: meta@dtb:uid content 'a8620173-6f09-4947-99a6-002454099f47' should
conform to unique-identifier in content.opf:
'urn:uuid:a8620173-6f09-4947-99a6-002454099f47' in OEBPS/toc.ncx"

I'm assuming that this is a Table of Contents error, but can anyone advise how to fix it in either Sigil or InDesign?

Or is this just a case of finding the UID content entry with a text/HTML editor and changing it to read uuid:a8620173-6f09-4947-99a6-002454099f47? And if so, where should I look?

Advice appreciated! Thanks!

[Toxaris]

I assume you did some manual changes in either the content.opf and/or toc.ncx. Perhaps some copy/pasting of another ePUB.

Anyway, an ePUB must have an unique indentifier and this is specified in the content.opf. You could try to regenerate the TOC (not the HTML TOC) to fix this. This should solve it, as the toc.ncx should not contain an UUID.

[GrannyGrump]

Re-generating the toc is simpler, but if you have a special uuid, you can edit it manually. (You can edit the opf and the tox.ncx files with a text editor, or in the Sigil code window, but be very cautious)

I use Sigil, and I happen to create my own uuid manually, but even if I use the one generated automatically by Sigil, it will show in the content.opf as <dc:identifier blah blah ...

So the entry in the opf file looks like:

<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<package xmlns="http://www.idpf.org/2007/opf" unique-identifier="BookId" version="2.0">
<metadata xmlns:dc="http://purl.org/dc/elements/1.1/" xmlnspf="http://www.idpf.org/2007/opf">
<dc:identifier id="BookId" opf:scheme="UUID">urn:uuid:ecm-ABP-ST-24-07-2016</dc:identifier>

In the toc.ncx, the <meta content tag must match the opf file, thus

<meta content="urn:uuid:ecm-ABP-ST-24-07-2016" name="dtb:uid"/>
<meta content="3" name="dtb:depth"/>
<meta content="0" name="dtb:totalPageCount"/>
<meta content="0" name="dtb:maxPageNumber"/>

The problem with a mis-matched uuid is that saving the epub in that condition will result in Sigil trying to "repair" the OPF file and you end with rather complex "nested" metadata tags.

[DiapDealer]

I run into retail epubs all the time whose uid in the ncx doesn't match the opf uid. I match them up when I notice the discrepancy, but I have to seriously wonder if it really matters one single bit in the "real world."

When using Sigil, don't be afraid to alter the NCX dbt:uid value to match the OPF's uid. It's getting the OPF wrong that can wreak havoc.

[pstjmack]

Thanks, all. Still working on it!

[pstjmack]

Now, unfortunately, the Contents problem has sorted out, but a fresh one has cropped up. The file gets an error that says "ERROR: Unable to process encrypted EPUB".

I presume this is a problem either with InDesign implementing DRM unexpectedly, or with embedded fonts, which seem to cause the same error elsewhere. Does anyone have any ideas or suggestions on how to fix this error? Thanks!

[Toxaris]

This could very well be related to embedded fonts. I seem to recall that if you embed fonts with ID that you need to specify that you do not want to encrypt fonts.

[KevinH]

The unique id is used to encode/decode embedded fonts. Two different font encryption approaches exist (one Adobe's and one Idpf) and exactly how they encrypt/decrypt and exactly what part of the unique id string they use also seems to be not well specified so there are variations on that as well.

For Adobe's version I would remove the "urn:uuid:" prefix from the unique id in the content.opf and the toc.ncx (making sure they match) as I think only the uuid part is used. For idpf, they use whatever is provided in that unique id , leading to font encryption vs decryption issues depending on how these routines were implemented across devices.

Hope something here helps.

KevinH