PDF formatting tweaks

c) Numbered lists. The number is followed by a . but the . gets swallowed by the text if there are steps numbered with two digits (10, 11, ...). Topic "Adding basic fields to a content type" is an example.

I've made a small tweak to the PDF stylesheet file scripts/pdf.xls, which takes care of 2 of the 3 problems (b and c).

It turns out that the extra vertical whitespace between topic headers and the first text in the chapter is due to the index entries. We have things like:

[role="summary"]
Overview of Drupal and the reasons to choose it as a reliable Content Management System (CMS).

(((CMS (Content Management System),overview)))
(((Content Management System (CMS),overview)))
(((Drupal content management system,overview)))
(((Drupal content management system,server requirements)))
(((Drupal core,overview)))
(((Drupal.org website,downloading modules and themes from)))

//==== Prerequisite knowledge

==== What is a Content Management System?

The block of index entries is creating a paragraph in the output. It renders as a blank paragraph. This is true in the e-book as well as HTML output.

What we need to do to fix this is attach the index entries in the source to particular paragraphs of text. So it should be something like:

==== What is a Content Management System?

(((CMS (Content Management System),overview)))
(((Content Management System (CMS),overview)))
(((Drupal content management system,overview)))
A content management system (CMS) is a software tool that lets users add,
publish, edit, or remove content from a website, using a web browser on a
...

And then the other index entries would go with text about them.

There may be a way to fix this using a script... but not today. So, leaving this open.

Well, maybe I will get to it today. As a first step, I refactored the addnames script so that it can handle adding more preprocessing steps, and that a command line argument can decide which steps we need (we don't do it for ebooks). Next I can try to figure out a way to move the index entries to a better place reliably...

I managed to tweak the XSLT so that this is now working well for HTML and PDF/eBook output.