Check Book repair function has unexpected side-effect

DrChiper · 07-29-2015, 09:41 AM

Today, while fixing some ebook with the Edit Book function, I got the following warning on using the Check Book function:

Code:

WARNING: Invalid id: stärt [OEBPS/metadata.opf]

So I inspected the metadata.opf and found the following:

Code:

<manifest>
    <item href="Text/start.xhtml" id="stärt" media-type="application/xhtml+xml"/>

Check Book explained this as follows:

Code:

The id stärt is not a valid id. IDs must start with a letter ([A-Za-z]) and may be followed by any number of letters, digits ([0-9]), hyphens ("-"), underscores ("_"), colons (":"), and periods ("."). This is to ensure maximum compatibility with a wide range of devices.

Apparently the "ä" caused some problem (it is a valid letter though), so I tried to repair it with these two options of Check Book:

Replace this id with a randomly generated valid id
Try to correct all fixable errors automatically

The "repair" in both cases turned out to be the complete removal of the text file where ID stärt refers to.

I admit, the removal "cures" the problem, but I do not think that this is the right solution

In particular when you at first are not aware of it and later on seem to have a corrupted e-book with a missing chapter.

theducks · 07-29-2015, 10:34 AM

accented characters are not in the basic ASCII [A-Za-z]

someone got cute with the id

and made it invalid

DrChiper · 07-29-2015, 11:40 AM

Well, not quite. The original chapter text file name also contained that accented character, so I assume that that was also used to create the ID.

However, my major concern is the ruthless removal of that same chapter text file by Check Book to "cure" the mistake. Coincidentally I noticed it (the files chapter numbering had suddenly a gap) before pressing "save".

theducks · 07-29-2015, 12:05 PM

Quote:

Originally Posted by DrChiper

Well, not quite. The original chapter text file name also contained that accented character, so I assume that that was also used to create the ID.

However, my major concern is the ruthless removal of that same chapter text file by Check Book to "cure" the mistake. Coincidentally I noticed it (the files chapter numbering had suddenly a gap) before pressing "save".

definitely issues

Getting rid unsafe (filename) characters is one thing.
Loss of data (file) is not good

Your guess on the ID names is probably good,

and a good example why blind faith ID generation is not a good idea.

kovidgoyal · 07-29-2015, 12:07 PM

It should definitely not be removing the file, it should be replacing the id with a randomly generated safe id. It's probably because the id in question was in the OPF. Open a bug report and attach the file, I will take a look.

DrChiper · 07-29-2015, 12:16 PM

Quote:

Originally Posted by theducks

Getting rid unsafe (filename) characters is one thing.

I checked, but using accented characters like "ä" in xhtml file names does not cause calibre to issue any warning, like for instance spaces certainly will do.

Incidentally, the original ebook was not produced by calibre.

DrChiper · 07-29-2015, 01:18 PM

Submitted Bug #1479462 + attached test file

theducks · 07-29-2015, 03:11 PM

Quote:

Originally Posted by DrChiper

I checked, but using accented characters like "ä" in xhtml file names does not cause calibre to issue any warning, like for instance spaces certainly will do.

Incidentally, the original ebook was not produced by calibre.

Not using some things in a book is a best practice, even if EPUB says it is OK
All those lame render engines out there, are the reasons to stay with KISS

DrChiper · 08-07-2015, 08:50 AM

This bug is indeed fixed now in version 2.34
Thanks Kovid

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Unexpected Book downloaded without purchase/request	eddomak	Kobo Reader	24	07-11-2014 03:52 AM
Name A Book (if any) that had a profound effect on you.	Marcus365	General Discussions	43	03-13-2014 12:26 PM
902 Comic book / PB902 side by side comparison	Lester Burnham	PocketBook	26	10-07-2011 03:35 AM
Weird sunlight side-effect on my K2	delphidb96	Amazon Kindle	5	08-20-2010 08:54 PM
Can you post a list of functions of the bottoms (icons) on the e-ink side here?	henry_moh	enTourage Archive	4	05-10-2010 12:55 AM

07-29-2015, 09:41 AM	#1
DrChiper Bookish Posts: 1,017 Karma: 2003162 Join Date: Jun 2011 Device: PC, t1, t2, t3, Clara BW, Clara HD, Libra 2, Libra Color, Nxtpaper 11	Check Book repair function has unexpected file remove side-effect Today, while fixing some ebook with the Edit Book function, I got the following warning on using the Check Book function: Code: WARNING: Invalid id: stärt [OEBPS/metadata.opf] So I inspected the metadata.opf and found the following: Code: <manifest> <item href="Text/start.xhtml" id="stärt" media-type="application/xhtml+xml"/> Check Book explained this as follows: Code: The id stärt is not a valid id. IDs must start with a letter ([A-Za-z]) and may be followed by any number of letters, digits ([0-9]), hyphens ("-"), underscores ("_"), colons (":"), and periods ("."). This is to ensure maximum compatibility with a wide range of devices. Apparently the "ä" caused some problem (it is a valid letter though), so I tried to repair it with these two options of Check Book: Replace this id with a randomly generated valid id Try to correct all fixable errors automatically The "repair" in both cases turned out to be the complete removal of the text file where ID stärt refers to. I admit, the removal "cures" the problem, but I do not think that this is the right solution In particular when you at first are not aware of it and later on seem to have a corrupted e-book with a missing chapter. Last edited by DrChiper; 07-29-2015 at 01:21 PM.

07-29-2015, 10:34 AM	#2
theducks Well trained by Cats Posts: 31,057 Karma: 60358908 Join Date: Aug 2009 Location: The Central Coast of California Device: Kobo Libra2,Kobo Aura2v1, K4NT(Fixed: New Bat.), Galaxy Tab A	accented characters are not in the basic ASCII [A-Za-z] someone got cute with the id and made it invalid

07-29-2015, 11:40 AM	#3
DrChiper Bookish Posts: 1,017 Karma: 2003162 Join Date: Jun 2011 Device: PC, t1, t2, t3, Clara BW, Clara HD, Libra 2, Libra Color, Nxtpaper 11	Well, not quite. The original chapter text file name also contained that accented character, so I assume that that was also used to create the ID. However, my major concern is the ruthless removal of that same chapter text file by Check Book to "cure" the mistake. Coincidentally I noticed it (the files chapter numbering had suddenly a gap) before pressing "save".

07-29-2015, 12:07 PM	#5
kovidgoyal creator of calibre Posts: 45,351 Karma: 27182818 Join Date: Oct 2006 Location: Mumbai, India Device: Various	It should definitely not be removing the file, it should be replacing the id with a randomly generated safe id. It's probably because the id in question was in the OPF. Open a bug report and attach the file, I will take a look.

07-29-2015, 01:18 PM	#7
DrChiper Bookish Posts: 1,017 Karma: 2003162 Join Date: Jun 2011 Device: PC, t1, t2, t3, Clara BW, Clara HD, Libra 2, Libra Color, Nxtpaper 11	Submitted Bug #1479462 + attached test file

08-07-2015, 08:50 AM	#9
DrChiper Bookish Posts: 1,017 Karma: 2003162 Join Date: Jun 2011 Device: PC, t1, t2, t3, Clara BW, Clara HD, Libra 2, Libra Color, Nxtpaper 11	This bug is indeed fixed now in version 2.34 Thanks Kovid