[Booki-dev] Re : PDF Export

[Tuukka Hastrup]

2011-12-12 23:03, Nicolas Dufour kirjoitti:
>> De : Cedric Gémy<cedric.gemy at gmail.com> There might be some
>> settings to change between PDFtk and other used tools in the export
>> process.
>
> To add a bit to the weirdness of the situation, according to the PDF
> properties of our documents, it seems that we have three different
> PDF generators.
> 1. When the PDF is generated from the export tab of
>    booki, we have: iText 2.1.7 by 1T3XT/pdftk 1.41 (with potential
>    license issues with iText, which is considered commercial because we
>    use it to serve PDFs on the fly in a web application, see
>    http://itextpdf.com/terms-of-use/index.php).
> 2. When the PDF is generated from Objavi: GPL Ghostscript 8.71.
> 3. When imported from Booki Publisher: wkhtmltopdf.
>
> I thought everything was linked to Objavi and thus used the same
> converter. Surprisingly it's not the case. Why?

Objavi uses multiple commands in its html-to-pdf processing pipeline:

* wkhtmltopdf to convert the content from html to pdf
* pdfnup to make multi-column pdf
* pdfedit script wk_objavi.qs to post-process the pdf
* bookland and ps2pdf to create a barcode
* pdftk to rotate a pdf
* pdftk to combine the book part pdfs into one final pdf
* gs to embed all fonts in the pdf

Not all of these are used always (depending on the settings and the 
exact pipeline used). I suppose that's why the PDF generator of the 
final file varies.

> Adam, the Objavi source code hosted in flossmanuals.net is quite
> outdated, and the one on GitHub is empty (only a README file). Could
> you please tell me where I can find the most recent files?

I think the latest published version is still at

git://booki-dev.flossmanuals.net/git/objavi2.git

We have something of a resource problem in cleaning up and documenting 
things. The Objavi git repo is pending on someone to rewrite its history 
to purge some historical files before pushing the repo to GitHub.