Help with validation error(s): "reacheable" not in <manifest> + others

[bubism]

Hello,

I am here to humbly ask the wisdom of this forum about the last validation error appearing in my .epub. Can anybody help?

THE CONTEXT
The book is a heavily (about 10K) internally hyperlinked dictionary, with lots of anchors and internal links.

THE ERROR (within Sigil)
OEBPS/Text/ N/A "This resource is reachable but not present in the OPF <manifest>. "Reachable" means that a reference of some kind that points to this resource exists in the epub."

THE ERRORS (with Epub validator)
Lots of errors unreported by Sigil, all of them of two types:
1) 'OEBPS/Text/': referenced resource missing in the package.
2) 'placeholderword': fragment identifier is not defined in 'OEBPS/Text/XXX html_split_001.html'

MY UNDERSTANDING
I guess that the Epub Validator errors are simply caused by the automatic segmentation of the epub file. My understanding is that these errors are reported whenever an internal link points to an anchor outside of the same segment

What worries me is the single Sigil error, because the <manifest> list seems to contain all the file references, and of course I cannot manually check 10K+ links without even a hint of where the error seems to be located.


MY QUESTIONS
a) Does anybody know how to pinpoint the code line causing the Sigil error?
b) Are the Epub Validator errors anything to worry about, and in case how can they be corrected?



Thank you in advance for your kind help... I'm keeping my fingers crossed!

[theducks]

Some errors show a line number where the error became apparent.
Just doubleclick on the filename next to the error: Presto

Now to discuss apparent:
Something happened way back there somewhere, like you opened a <div tag , but it never closed and we now see a </body

relative links can become broken with section splits:
1) Just avoid using the use, even if it is legit
2)REGEX search for: Text\/\# (Finds: Text/#fooeyonus. anchors without a filename)

[bubism]

Thank you, Theducks, but I'm afraid your answer doesn't apply to my issue.

Specifically:

the very problem of the Sigil error is that no line number is given

and

searching for "Text\/\#" obviously finds every internal link, but there is no way to distinguish the 10k working ones from the single (possibly) malformed culprit.


(I restructured the file merging back the sections and recreating the TOC, but the errors remain)

[PeterT]

Maybe some screen prints might help folks see what might be happening.

[theducks]

Originally Posted by bubism

searching for "Text\/\#" obviously finds every internal link, but there is no way to distinguish the 10k working ones from the single (possibly) malformed culprit.

You have 10K broken (relative?) links
text/# only finds relative links.

You need the ID part and if you reused the same id on different pages
(there are things that are Valid, but best not to use. Unique )

[Toxaris]

I don't how you created all the links, but at least some of them will not work. These are real errors. You have references in your book that are using wrong names.

Check your workflow with regards to these links.

[st_albert]

Also, beware of filenames that contain spaces or other funky characters.

[bubism]

Update

The links were created within Dreamweaver. It looks like the HTML to EPUB conversion (with Calibre) randomly deleted several anchors. I am now reinserting them in Sigil and the errors are slowly disappearing.

Other errors were caused by missing extended references to the right split where the anchors were. Since we are talking about roughly 1,700 anchors here, the only possible solution was to merge all the text into one file, but I am now very worried by how the readers will manage one sizeable file (310K).

Any suggestions will be very appreciated.

[PeterT]

Originally Posted by bubism

It looks like the HTML to EPUB conversion (with Calibre) randomly deleted several anchors.

Might want to open a calibre bug report and include the files in question.

[st_albert]

Originally Posted by bubism

Update

The links were created within Dreamweaver. It looks like the HTML to EPUB conversion (with Calibre) randomly deleted several anchors. I am now reinserting them in Sigil and the errors are slowly disappearing.

Other errors were caused by missing extended references to the right split where the anchors were. Since we are talking about roughly 1,700 anchors here, the only possible solution was to merge all the text into one file, but I am now very worried by how the readers will manage one sizeable file (310K).

Any suggestions will be very appreciated.

If it were me, and if possible, I would go back to the DreamWeaver HTML output and import it into Sigil rather than go through Calibre. Once in sigil, you can split it into reasonable "chapter files" (it will be slow working until you do this) and all the links should be correctly handled. At least, for me they are, but I've never tried 1700 anchors at a time.

What I'd do, before splitting, is go through and add sigil split markers where you want them (perhaps can be done via search-and-replace), and maybe do other things that make sense to be done over the whole book (e.g. add chapter headers or the like). Then split the book, validate, and clean up as necessary.

ETA: Also, before splitting, be sure to convert any in-line css definitions (located in the <head> section of the html file) to a separate .css file in sigil. That way, you only have to do it once, and the CSS will all be in one place rather than in each of the split chapters, for easy maintenance. (please ask if you need further instructions or clarification on this -- but do it for sure.) I don't work with dream weaver, but in the best of all possible worlds it would already produce a separate .css file in its output. Then just import that into the "Styles" directory in Sigil, make sure the appropriate link is provided in the <head> section of the HTML file, and you're good to go.

This is my normal workflow when dealing with epub export from InDesign CS4, which outputs the entire book in one .xhtml file. The largest I've worked on so far contained ~ 18,000 lines of code. It was a bit sluggish, but workable with the previewer turned off. (I usually work exclusively in code view.) So that's how I'd be inclined to handle your situation.

I've nothing against Calibre, but it is not known for producing especially human-readable code, and besides, from what you're saying it looks like this situation is beyond its capablilities.

BTW, the "sigil split marker" I'm referring to is

<hr class="sigil_split_marker" />

and is inserted (in code view) at the point where you want the split to occur. For me, that is usually just before the next chapter header.

Of course, you can insert them via book view as well, but you have to be very careful how you do it if you want to get the expected result. YOYO in that case, since I never do that. Other threads here (sigil forum) will have good information.

HTH

Albert

[Notjohn]

If I insert 20 or 30 "Sigil split markers" in my html file, then open the html in Sigil, will it be one file or many?

Sounds like a time-saver!

[DiapDealer]

Originally Posted by Notjohn

If I insert 20 or 30 "Sigil split markers" in my html file, then open the html in Sigil, will it be one file or many?

It will be one file. Right up until such time as you select "Split at Markers" from the Edit menu (F6 default keyboard shortcut). Then it will become many.

[bubism]

@st albert - thank you for the info. I am afraid it is too late now to go that way, but I will surely adopt your workflow for my next books (no more dictionaries, thank god!).

I'll let you know how it ends...

[DiapDealer]

In the future (especially if the book is quite complicated/complex), I'd also recommend not leaving validation as the very last step (which I fear we're all probably a liitle guilty of). If you have tons of links, it's probably a good idea to validate regularly--but especially before and after splits/merges or bulk s&r. Things are always easier to fix (roll back) when you know exactly what you were doing when things went pear-shaped.

I know that's not very helpful right now… and I apologize for that, but sometimes there's just no help to be had for "needle in a haystack" type situations like this.

[st_albert]

Originally Posted by DiapDealer

It will be one file. Right up until such time as you select "Split at Markers" from the Edit menu (F6 default keyboard shortcut). Then it will become many.

And, before you split it, the split points will show up in book view or the previewer as a horizontal line indicating the split point. (which is what the <hr... /> code does)