This should have been an established process, but for some reason it was
edited manually after initial generation.
After the change `reference.xml` can be removed from the repository, but
I left it to show that the things are working fine and the difference from
an auto-generated one are simple divergences due to manual changing
(not updated `reference.xml` with added `<purpose>` elements in 38017203f69cec3aa5f402fe8306eb1deff67b9a). Also removed the date numbers
from `dtdxml` as they were never in `xml` and seems to have no actual use,
but brings a lot of noise to `xml` differences after added auto-generation.
In the last commit I changed it to store in fast-elements with the wrong id.
Unfortunately this can't be picked up in testing, as the fallback still works
correctly. Although it didn't make any notable difference to the build time,
so maybe the fast-elements mechanism isn't needed?
I get HTTP 303 errors when retrieving images from the old location (browser is fine, but Java chokes while building the PDF's). I couldn't find a git location that didn't also give 303 errors, so this seems to be the next best alternative.
Whatever version of xsltproc I'm not using generates even longer id
numbers, which means that the base id needs to be further truncated to
make space for them.
It turns out the '\code' blocks don't have this problem, so an easier
fix would be just to use them. But now I've implemented the workaround,
might as well leave it in, in case anyone wants to use <pre> blocks in
the future.
This improves layout on mobile devices. I had to remove declaration
for the 'rev' element as otherwise the 'meta' element would have
namespace declaration as well, for no reason. Possibly there's a
better solution.
The footer template has code to generate revision number,
which was never updated to git and therefore generates no text,
but still generates a somewhat complex table structure. This
commit removes code for version numbers, and makes output a simple
div.
I've also added a hook to use pick content of whatever paragraph
with role of 'copyright' the document has, since right now we
pick copyright only from 'bookinfo' element, and Docbook renders
bookinfo element right at the start, and for Boost.Build documentation
I want documentation to start with some hints how to start, not
with a copyright statement.
Remove stray text on end of line.
Fixes PDF doc build.
Aside: this whole Q&A set block could probably be removed now given that it appears to be a workaround for the ancient FOP-0.2.
