Hi, I started migrating the 5.0 docbook documentation to asciidoctor, and, because there wasn't any converter for such a task, I wrote a small shell script which uses pandoc. I managed to generate the asciidoc files from the current docbook, by using the html format the docbook generates. The resulting asciidoc files need some minor changes, but all in all it's still better than doing it manually. I wrote a new gradle task to take the asciidocs from the mapping guides and generate the html documentation, while applying the Hibernate CSS. I saw there are some common XSLTs that were applied in order to add the Hibernate logo and the Copyright info to every HTML page. If you want to take a look on it, this is my branch with the latest changes I've made: https://github.com/vladmihalcea/hibernate-orm/tree/feature/asciidocs >From a discussion I found on the asciidoctor GitHub page ( https://github.com/asciidoctor/asciidoctor-gradle-plugin/issues/52 ), I realized that we could create a tool-chain like this: - we write all docs in asciidoctor and use the asciidoc-gradle-plugin to convert those to docbook - from docbooks we can use the current jdocbook plugin to generate the single HTML, multi-html and PDFs and use the current XSLTs. With this workflow, the styling of the docs is already resolved, since the jdocbook configs can take care of it already. This could be a first phase until we figure out how we can generate everything we want and customize it all without having the intermediate docbook phase. Or, we can just forget about docbook and do our best to generate everything using the asciidoctor tool-chain, even if it might not be a straightforward task. Let me know what you think? Vlad _______________________________________________ hibernate-dev mailing list hibernate-dev(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/hibernate-dev

Hi, On Wed, Dec 02, 2015 at 01:28:59PM +0200, Vlad Mihalcea wrote: > I started migrating the 5.0 docbook documentation to asciidoctor, and, > because there wasn't any converter for such a task, I wrote a small shell > script which uses pandoc. There is a converter - https://github.com/bleathem/docbook2asciidoc That's the converter we used for the initial conversion of the Hibernate Search and Validator docbook sources to asciidoc. It works quite well. The output is directly usable. The actual generated asciidoc can need some cleanup which we did in a once off effort. AFAIR, there were a couple of caveats which we only discovered after the migration and after some manual changes. One of them was that we used in Search and Validator a custom docbook class to mark classes (maybe also method, not sure). The XLST in these scripts does not take care of this, even though it could with some minor tweaking. The result was that class names which used the 'classname' marker did not get special treatment in asciidoc. We ended up adding it manually. --Hardy