Sigil-0.9.3 Released

[KevinH]

It is dc:language in the metadata. You can add it once for the entire document. The language setting is required for both epub2 and epub3. The next version of Sigil will have a new epub metadata editor gui to make adding it even easier. Both Epubcheck and flightcrew will let you know if you miss setting it.

If you like both, then simply use search and replace on the html tag to add it to all html files in one quick command.

Hope this helps,

KevinH

[DiapDealer]

Originally Posted by roger64

I just had a look to the content.opf file of a new EPUB3 and this information seems to be missing as well. I have no personal problem about it, but this seems a bit curious. Who does not need any language?

You'd find the same thing in an EPUB2. Sigil as never pre-populated the dc:language metadata in a new epub's OPF. A validator will let you know you need to add it, though.

The gui metadata editor (for epub2) would default to Sigil's interface language (I think) if no dc:language metadata was present, but that's about it. You still had to open the editor and save it for changes to take effect in the OPF, though. Metadata is always going to be primarily up to the user to take care of.

[patrik]

(Please let me know if this should be in a separate thread.)

I just ran a (legally obtained) pdf through Finereader as I usually do with pdf's I like to read, but when I opened it with Sigil 0.9.3 I got the message:

This EPUB has HTML files that are not well formed. Sigil can attempt to automatically fix these files, although this can result in data loss.

There is one .xhtml file that it changed when saying "yes". Looking at the differences I'm not sure exactly what changed, besides many LF added making it look better in code view. (It seems to be exactly the same in book-view as far as I can tell with a quick glance.)

0.8.6 makes no error message when opening the epub-file. It looks even "better" in code view. (All xhtml-files seems to be.)

From an "easy to read" point of view, I prefer the way 0.8 does it.

Is this a beautify-thingy?


(Running FlightCrew both versions give 4 errors "Error schema not satisfied". Flightcrew in "fixed" 0.9 has an additional "Error schema not satisfied : element 'a' is not allowed for content model '(p|h1|h2|h3|h4|h5|h6|div|ul|ol|dl|pre|hr|blockquo te|address|fieldset|table|switch|form|noscript|ins |del|script)' near column 8".)

[KevinH]

The old 0.8.6 always ran the eqivalent of Tidy beautify that sometimes ate code badly and is not suitable for epub3 at all. Sigil 0.9.3 no longer does that. Sigil 0.9.3 uses a new parser that tries to change the code being sent in as little as possible. You can beautify the entire book if you want to still via the tools menu. None of these changes should have an impact on BookView (what another ebook reader would produce).

The errors are because Finereader or whatever you used did not produce technically correct xhtml that is required by the xhtml rules. Sigil 0.9.x will autofix those for you.

The errors from epubcheck/flightcrew indicate that the epub does not meet the official epub2 or epub3 standards and should be fixed. That error indicates that you are using an anchor someplace where it is not allowed. If you click on the error, it should bring you close to the error location so you can fix it.

KevinH

[DiapDealer]

The new version of Sigil doesn't automatically "beautify" code. What you're seeing is how the code is actually presented in the files. You can always manually Prettify the code by using Tools (from the menu)->Reformat HTML->Mend and Prettify All HTML Files. That will present the code in more of a format that you're used to seeing with the older version's Pretty-Print. You can also right-click on the Code View widget and select Reformat HTML from the context menu and Mend (or Mend and Prettify) the current file-only if you wish.

As for the error you're getting upon opening, I couldn't accurately say without seeing the actual code. My guess is that the older version of Sigil is simply fixing the error in the background with no warning message, where the new version is telling you about it.

EDIT: and what KevinH said.

[patrik]

Thanks guys, your support is beyond amazing. :-)

(I would be happy to send the pdf and the Finereader-created epub to you if you think that would help in any way.)


Finereader is fantastic when it comes to doing ocr. Nothing else I have tried and used even comes close. But for some reason it messes up some "simple" (in my uninformed view) things, like putting parts of a "CONTENTS"-page in the end of a xhtml-page way further in the book. Easy to fix, but kind of strange. :-)
The Flightcrew-errors are all bookmarks to such things.

[theducks]

PDF conversions are always strange (See the sticky in Calibre Conversions)
With luck, it is only a tiny bit strange
But I have seen Horrors

Be prepared to do cleanup

[Doitsu]

Originally Posted by roger64

I just had a look to the content.opf file of a new EPUB3 and this information seems to be missing as well. I have no personal problem about it, but this seems a bit curious. Who does not need any language?

@roger64: BTW, KevinH has already updated the Sigil code so that the next version will automatically insert a placeholder dc:title metadata entry and a dc:language metadata entry based on the Default Language for Metadata setting for both epub2 and epub3 files.

[DiapDealer]

Originally Posted by Doitsu

BTW, KevinH has already updated the Sigil code so that the next version will automatically insert a placeholder dc:title metadata entry and a dc:language metadata entry based on the Default Language for Metadata setting for both epub2 and epub3 files.

Hey, you're right! I need to pay closer attention.

[roger64]

Originally Posted by Doitsu

@roger64: BTW, KevinH has already updated the Sigil code so that the next version will automatically insert a placeholder dc:title metadata entry and a dc:language metadata entry based on the Default Language for Metadata setting for both epub2 and epub3 files.

I know Kevin is too good!! (I am not joking)

Bug corrections

I can confirm the correction of these two as they were part of 9.3. bug corrections:

1 - scrolling for a Linux user is now smooth as silk. This is quite a feat! (Archlinux - 64 bits. Sigil 9.3.1.)

2 - index: Sigil now takes care of locale for sorting the index entries according to each language alphabetical order. For French, it means that accented letters are not trailing anymore at the back of the common alphabet. Accented E is now placed just before the normal E (normally it should be placed just after).

[diegoan]

Hi there,
It detected an error in Sigil 0.9.3 on Windows 10 with F8 when the author's name is edited and placing two separated by & the authors clears out & leaving a single author.
You have to manually edit the conten.opf to put more than one author.
Cheers

[KevinH]

Each dc:creator is used for one author. Add multiple dc:creator or dc:contributer to add others. Did you enter the ampersand as an entity?

KevinH