Re: [docbook-apps] Docbook to epub - Extra field length for first filename must be 0, but was ..

[Carlos Araya <carlos.araya@gmail.com>]

Lars, one solution is to use the exec task and call the zip system file (I work on OS X and Linux, your mileage may vary)

<target name="epub-zip">
    <!-- We use exec zip because I can't figure out if the builtin zip task
    conforms to what ePub expects -->
    <!-- Zip the mimetype first without compression as required-->
    <exec executable="zip">
        <arg line="-X0 ${src_file}.epub mimetype"/>
    </exec>
    <!-- Then zip the other files into the zip/epub archive -->
    <exec executable="zip">
        <arg line="-r -X9 ${src_file}.epub META-INF OEBPS"/>
    </exec>
</target>

Hope this helps

On Mon, Feb 17, 2014 at 9:15 AM, Lars Vogel <lars.vogel@gmail.com> wrote:

Hi,

maybe the first line in my zip task was misleading, I created temp.mimetype. If I adjust it to use the file name you suggest it does not make a difference.

<zip destfile="${output.dir}/mimetype.zip" basedir="${output.dir}/tmp" compress="false" includes="mimetype" />

<zip destfile="${output.dir}/temp.zip" basedir="${output.dir}/tmp/" level="9" compress="true" excludes="mimetype" includes="OEBPS/** META-INF/**" />

<!-- Missing option to add -X as parameter -->

<zip destfile="${output.dir}/book.epub" update="true" keepcompression="true" encoding="UTF-8" excludes="*.html">

<zipfileset src="" />

<zipfileset src="" />

</zip>

I read in Bobs description for epub3 (http://sourceforge.net/projects/docbook/files/epub3/) that I need to specify -X for the Ant task to avoid this error. Is this something I can do with the Apache Ant task?

Best regards, Lars

2014-02-17 17:50 GMT+01:00 Nils Cordes <info@nilscordes.de>:

Hi Lars,

I had the same issue and fixed it that way.

See the comment of Matt Garrish at http://idpf.org/comment/370#comment-370

"When you zip your EPUB, the mimetype file must be the first one you add (mimetype == 8 characters)."

Ciao

Nils

Nils Cordes

Plesser Str. 11

12435 Berlin

Deutschland

Telefon: +49 176 10091551

E-Mail: info@nilscordes.de

Web: http://nilscordes.de

2014-02-17 17:33 GMT+01:00 Lars Vogel <lars.vogel@gmail.com>:

Hi Nils,

This is IMHI correct, I create it in the first step. Also this seems to be unrelated to the error message.

Best regards, Lars

Am 17.02.2014 17:28 schrieb "Nils Cordes" <info@nilscordes.de>:

Hey Lars,

I think there is an error in the zip part of your Ant buildfile. Try to change "temp.mimetype" to "mimetype.zip".

<zip destfile="${output.dir}/mimetype.zip" ...

<zipfileset src="" ...

Best regards

Nils

2014-02-17 16:32 GMT+01:00 Lars Vogel <lars.vogel@gmail.com>:

Hi,

I'm generating epub from Docbook via Apache Ant and I'm getting the following error during the validation with Epubcheck.

     [java] Epubcheck Version 3.0ERROR: /home/vogella/workspace/docu/de.vogella.publishing/../output/epub/book.epub: Extra field length for first filename must be 0, but was 20

Any suggestion what might be causing the error.

I generate the epub via the following snippet (see below).

Thanks for any help, Lars

<target name="build-epub_eclipsercp" description="Generates EPUB files from DocBook XML">

<xslt style="${epub.stylesheet}" extension=".html" basedir="${output.dir}/tmp2" destdir="${output.dir}/tmp">

<include name="*80_Book_Eclipse_RCP*/*book.xml" />

<param name="epub.stylesheet" _expression_="style.css" />

<param name="html.stylesheet" _expression_="${stylesheet}" />

<param name="section.autolabel" _expression_="1" />

<param name="chapter.autolabel" _expression_="1" />

<param name="show.comments" _expression_="0" />

<param name="section.autolabel.max.depth" _expression_="1" />

<param name="toc.section.depth" _expression_="1" />

<param name="section.label.includes.component.label" _expression_="1" />

<param name="make.valid.html" _expression_="1" />

<param name="html.cleanup" _expression_="1" />

<param name="chunker.output.encoding" _expression_="UTF-8" />

<param name="ignore.image.scaling" _expression_="1" />

<outputproperty name="indent" value="yes" />

<classpath refid="saxon.class.path" />

</xslt>

<copy todir="${output.dir}/tmp/OEBPS">

<fileset dir="OEBPS">

</fileset>

</copy>

<copy todir="${output.dir}/tmp/META-INF">

<fileset dir="META-INF">

</fileset>

</copy>

<!-- Don't know how to avoid genereation of "${destfilename}.html" by Saxon -->

<delete file="${output.dir}/tmp/book.html" />

<echo message="Generating book.epub" level="info" />

<!-- We create temporary zips so that minetype is the first one in the final zip  -->

<zip destfile="${output.dir}/temp.mimetype" basedir="${output.dir}/tmp" compress="false" includes="mimetype" />

<zip destfile="${output.dir}/temp.zip" basedir="${output.dir}/tmp/" level="9" compress="true" excludes="mimetype" includes="OEBPS/** META-INF/**" />

<zip destfile="${output.dir}/book.epub" update="true" keepcompression="true" encoding="UTF-8" excludes="*.html">

<zipfileset src="" />

<zipfileset src="" />

</zip>

<!-- Have to delete these directories would be nicer to place then in tmp output dir 

<delete dir="./OEBPS" />

<delete dir="./META-INF" />

-->

<!-- Make sure the epubcheck lib has a subfolder lib with saxon.jar and jing.jar in it

-->

<epub.check epub="book" />

</target>