FC and Sigil 0.5.3 ePUBcheck failure

[Hitch]

Gang:

We had something happen today that is a bit unnerving; we had some "coffee-table books" that we'd made a variety of ways, but in our last iteration, had to resize all the images, so we exploded the ePUB, batch-resized the dir, and re-asssembled the epub. Now, as it turns out, there were two detritus files hanging about, which ended up in the images dir. (Don't ask about that part--it's not what's important).

The resulting ePUB validated in both 0.5.3 and FC. However, when the book was in Transport to iBooks, it failed ePUBcheck. Here are the two errors:

Error ISBN *************: WARNING:NAME\4-4-2012\ISBN.epub: item (OEBPS/Text/Ch01Img09.html) exists in the zip file, but is not declared in the OPF file.
Error ISBN *************: WARNING: \NAME\4-4-2012\ISBN.epub: item (OEBPS/Text/TOC.bak) exists in the zip file, but is not declared in the OPF file.

What's in RED I've altered for privacy.

However, both Sigil and FC should have found these errors. We checked and rechecked and validated several times, and neither ever found these in the re-packed ePUB. Any thoughts? This seems to clearly be a bug of some kind--but I can't figure out what/why. The only thing I can think is that the process of repacking somehow negated FC's ability to "see" the errors, but that seems bollix.

@user_none--I can send the "validating" old ePUB to you if you want to look at it, it's monstrously large, 64.5mb. Advise if you want it sent for review.

Hitch

[JSWolf]

What I do to make sure I don't have any straggler files is delete the contents of the ePub ZIP except for META-INF and mimetype. Then I put in the rest of the ePub and I know there are no file left in that I've deleted.

[Hitch]

Wolfie:

You missed the point. I'm not asking for procedural advice--we know that the bookmaker should have not used a dir with the detritus in it. The question is, why did FC and Sigil 0.5.3 MISS it?

ETA: ePUBcheck 3.0b4 does find these errors.

Hitch

[mmat1]

Originally Posted by Hitch

ETA: ePUBcheck 3.0b4 does find these errors.

There are actually TONS of errors which epub-check finds and FC within Sigil misses. Especially issues with links. Is that new to you ?

[Hitch]

No, mmat1, it isn't. It is, however, the first time I've had an issue of this "size." I've never had a problem with FC standalone missing an error that would cause an ePUB to fail epubcheck. My point to user_none is simply this: if it's THIS far out of whack, then there isn't any point in using it.

Hitch

[user_none]

In Sigil it's not found because Sigil ignores unmanifested files while importing. When the EPUB is opened by Sigil the .BAK file is simply not there.

I'm still investigating why it's not working with FC standalone.

[Hitch]

Originally Posted by user_none

In Sigil it's not found because Sigil ignores unmanifested files while importing. When the EPUB is opened by Sigil the .BAK file is simply not there.

I'm still investigating why it's not working with FC standalone.

user_none:

And the first error?: OEBPS/Text/Ch01Img09.html

Not sure I understand why this wouldn't have been found?

Hitch

[user_none]

Originally Posted by Hitch

user_none:

And the first error?: OEBPS/Text/Ch01Img09.html

Not sure I understand why this wouldn't have been found?

Sigil didn't generate an error on this one for the same reason it didn't on the .BAK file. When you open an EPUB with Sigil it ignores manifested files. That one not being in the Manifest was never imported into Sigil when it opened the EPUB.

[DaleDe]

Originally Posted by user_none

Sigil didn't generate an error on this one for the same reason it didn't on the .BAK file. When you open an EPUB with Sigil it ignores manifested files. That one not being in the Manifest was never imported into Sigil when it opened the EPUB.

Therefore if the file is saved it would be ok as none of these files would exist in the saved version. True?

Dale

[user_none]

Originally Posted by DaleDe

Therefore if the file is saved it would be ok as none of these files would exist in the saved version. True?

True. After saving in Sigil that particular EPUB would be valid.

[JSWolf]

So for now, would it be best to double check FC by running the ePub through ePubCheck after FC says all is well?

[Hitch]

Originally Posted by user_none

True. After saving in Sigil that particular EPUB would be valid.

NO. She opened the repacked ePUB, tweaked it IN SIGIL and saved it. And then it promptly passed FC and failed transport at Apple, and failed ePUBcheck.

Hitch

[eping]

This is a case that better ones can't win bad ones.
I think Sigil's Flight Crew and Threepress's PreFlight are both better and practical than ePubCheck, but since ePubcheck is released by the committee and considered as standard, FC and PF are embarrassed.
My opinion, ePub check is unfriendly, confusing, trivial, and even ugly. It even reports error that eReaders can support without any problem,but accept some that no eReaders support.

[mmat1]

Originally Posted by eping

My opinion, ePub check is unfriendly, confusing, trivial, and even ugly.

No!
It finds errors, Sigils built in FC don't. That's very friendly and nice ! The commandline interface is no problem to somebody, who has ever worked on an unix plattform (or Ms-Dos...). I understand very clear every massage (yes, epubmaking requires some experience...).

It even reports error that eReaders can support without any problem,but accept some that no eReaders support.

It reports violations to the standard. How can you be sure, that any ereader out in the world will tolerate them ? And if something is missing: well nothing ist perfect ...

You actually don't say, why you find epubcheck that worse.

I say, using it won't hurt.

[user_none]

FlightCrew does not currently check for un-manifested files. That's why it says it found no issues.

Originally Posted by OCF 2.0.1

Any extra files in the ZIP archive (i.e., files within the ZIP archive that are not listed within the package files’ <manifest> element, such as META-INF files or alternate derived renditions of the publication) MUST NOT be used in the processing of the OPS publication.

From the looks of it the EPUB 2 spec is saying un-manifested files should just be ignored and they are not errors.