Understanding two reported mistakes

[roger64]

Hi

I am just beginning to use the book-editor with Linux and have been very pleasantly impressed by its recent huge progresses.

I just checked one EPUB2 of mine and I have two questions about reported errors:

the first one (in French) is

Le fichier META-INF/com.apple.ibooks.display-options.xml [META-INF/com.apple.ibooks.display-options.xml] n'est pas compris dans le manifeste

This error has not been automatically corrected, thus my question:

Question: which text should I add in the manifest and where exactly to insert it?

the second one (concerning many xhtml files)

Entités nommées actuelles

(about named entities)

The reported mistakes have been all automatically corrected. However this correction is a kind of black box for me.

Question: Is there somewhere a kind of transposition table which explains that all A_entities have been changed to B and so on?

If I wish for some reason to come back to the previous state, I feel unable to do it as I do not know what has been changed and how important it was. Was it one entity or one thousand?

[theducks]

Originally Posted by roger64

Hi

I am just beginning to use the book-editor with Linux and have been very pleasantly impressed by its recent huge progresses.

I just checked one EPUB2 of mine and I have two questions about reported errors:

the first one (in French) is


This error has not been automatically corrected. thus my question:

Question: which text should I add in the manifest and where exactly to insert it?
the second one (concerning many xhtml files)

Entités nommées actuelles

(about named entities)

The reported mistakes have been all automatically corrected. However this correction is a kind of black box for me.

Question: Is there somewhere a kind of transposition table which explains that all A_entities have been changed to B and so on?

If I wish for some reason to come back to the previous state, I feel unable to do it as I do not know what has been changed and how important it was. Whas it one entity or one thousand?

Beautify all files will usuall fix the second. All it does is replace 'Named' entities with their numbered counterpart. & nbsp; becomes & #160;

There are 2 kinds of messages (at least that is all I have seen). Marked:
1) with a X icon. These are bad, If you select, suggestions appear in the right pane.

2)Yellow Warnings: These are items that don't belong to the spec / misbehave in some cases . They usually only mater depending on who gets the files. Brain dead validation at some major sites will require that even these must be resolved.

[roger64]

Thanks for your reply. Up to now I only have these -yellow- warnings quoted above.

I use a lot of &#-82-39; (without the two - in the middle). These are narrow no-break spaces

If I understand you correctly these last ones are left untouched?

[Doitsu]

Originally Posted by roger64

I just checked one EPUB2 of mine and I have two questions about reported errors:

the first one (in French) is

This error has not been automatically corrected, thus my question:
Question: which text should I add in the manifest and where exactly to insert it?

AFAIK, none of the files in the META-INF folder need to be manifested; the contents of this folder is ignored by epubcheck anyway, which seems to be one of the reasons that Apple puts all their custom .xml files in this folder.

I.e., technically speaking, Calibre Editor, is correct about reporting unmanifested files, because META-INF isn't supposed to contain files other than container.xml, but you can ignore this warning.

[mrmikel]

You will see non-breaking spaces disappear. They are translated into unicode rather than remaining in HTML. But they are there. You can see them at the lower right of the window as you cursor through the text.

[eschwartz]

most entities will be transformed into their unicode equivalent. Only if it would interfere with (x)html parsing will it retain the entity, and I don't think it will be changed from named to numbered.

Sigil is the one that converts named entities to numbered entities.

[roger64]

Originally Posted by mrmikel

You will see non-breaking spaces disappear. They are translated into unicode rather than remaining in HTML. But they are there. You can see them at the lower right of the window as you cursor through the text.

When I use the cursor, I can read as you say in the lower right part the name of the entity. The calibre editor identifies this way by their own names the no-break space (displayed by a yellow square in code view) and the narrow no-break space (displayed by a plain blank in code view).

However, the editor does not seem to change the nbsp; to #160. After saving the calibre correction, I opened again the EPUB file in sigil and I still had all my 1077 nbsp; like before.

I think there is a need to publish somewhere a kind of transposition table to explain clearly what changes will be (or should be) performed by the calibre editor and on which entities.

Originally Posted by Doitsu

AFAIK, none of the files in the META-INF folder need to be manifested; the contents of this folder is ignored by epubcheck anyway, which seems to be one of the reasons that Apple puts all their custom .xml files in this folder.

I.e., technically speaking, Calibre Editor, is correct about reporting unmanifested files, because META-INF isn't supposed to contain files other than container.xml, but you can ignore this warning.

OK thanks for this clarification.

[eschwartz]

Originally Posted by roger64

When I use the cursor, I can read as you say in the lower right part the name of the entity. The calibre editor identifies this way by their own names the no-break space (displayed by a yellow square in code view) and the narrow no-break space (displayed by a plain blank in code view).

However, the editor does not seem to change the nbsp; to #160. After saving the calibre correction, I opened again the EPUB file in sigil and I still had all my 1077 nbsp; like before.

I think there is a need to publish somewhere a kind of transposition table to explain clearly what changes will be (or should be) performed by the calibre editor and on which entities.

As I said in my previous post, calibre does not change named entities to numbered entities. It changes both to their unicode equivalent, either when you type the entity, or when you run Beautify/Fix HTML on the containing file. Additionally, Check Book will warn you of named entities (but not numbered ones) and offer to fix them.

In short, he following changes are performed:
named entities --> unicode --- when typed in, or fix/beautify files, or in Check Book
numbered entities --> unicode --- when typed in, or fix/beautify files

Exceptions are < > which will never be changed since that would mess up html.

° ' (and possibly more) appear to not change when typed in, but even those get replaced with Fix HTML.

[roger64]

Originally Posted by eschwartz

As I said in my previous post, calibre does not change named entities to numbered entities. It changes both to their unicode equivalent, either when you type the entity, or when you run Beautify/Fix HTML on the containing file. Additionally, Check Book will warn you of named entities (but not numbered ones) and offer to fix them.

In short, he following changes are performed:
named entities --> unicode --- when typed in, or fix/beautify files, or in Check Book
numbered entities --> unicode --- when typed in, or fix/beautify files

Thanks for your detailed answer but I fail to reproduce it. I am more than probably dumb but what happens is this: I open one EPUB2 of mine with the calibre editor and I ask it to correct and beautify this EPUB2. I save it. I open the "corrected" EPUB2 later with Sigil and I can still find inside over 900   living a happy and undisturbed life

Nothing is fixed on this regard. What am I expected to find?

[eschwartz]

Originally Posted by roger64

Thanks for your detailed answer but I fail to reproduce it. I am more than probably dumb but what happens is this: I open one EPUB2 of mine with the calibre editor and I ask it to correct and beautify this EPUB2. I save it. I open the "corrected" EPUB2 later with Sigil and I can still find inside over 900   living a happy and undisturbed life

Nothing is fixed on this regard. What am I expected to find?

If you re-open it in Sigil, Sigil will convert the unicode back into an entity.

[kovidgoyal]

Do not ever use Sigil to check what changes have been made to the book. Sigil modifies the book just by opening it.

Either use the calibre compare book tool or use the calibre unpack book tool to take a look at the contents of the file. Or use the editor itself, since it does not make any changes to the html when simply opening the file.

[roger64]

Originally Posted by kovidgoyal

Do not ever use Sigil to check what changes have been made to the book. Sigil modifies the book just by opening it.

Either use the calibre compare book tool or use the calibre unpack book tool to take a look at the contents of the file. Or use the editor itself, since it does not make any changes to the html when simply opening the file.

Thanks both of you for your replies. Now all this makes sense. So this nice calibre editor must absolutely be the last and final step in my workflow.

I tried the "compare" tool which is very handy. This is discovery day!!

I saw what happens to &nbsp and nnbsp (see entities snapshot and red squares) and how doctype is silently killed! Utterly frightening!