[cairo] Clickable links in PDF? Metadata?

[David Kastrup]

Khaled Hosny <khaledhosny at eglug.org> writes:

> On Fri, Jul 03, 2015 at 06:37:10PM +0200, David Kastrup wrote:
>> 
>> Hi,
>> 
>> I'm currently taking a look at the suitability of using Cairo for the
>> rendering purposes of GNU LilyPond <URL:http://www.lilypond.org>.
>> Hardly surprising, print output is pretty much a priority.
>
> SILE[1] was using Cairo as its PDF backend, but the final conclusion, I
> think, was is that it is not suitable for the job and now it uses a PDF
> library based on dvipdfmx[1].
>
> Regards,
> Khaled
>
> 1. https://github.com/simoncozens/sile
> 2. https://github.com/simoncozens/dvipdfm-x

Well, SILE appears to be much more text-heavy (sort of the Pango
domain).  LilyPond has lots of graphical content placed with drawing
instructions, and there is a strong case for a large variety of
reasonably fast backends that is not really satisfactorily covered by
PDF.

While it may be conceivable to recode our existing PostScript-producing
code into producing PDF (we don't lose any of the Ghostscript-based
conversions since Ghostscript will read PDF just as well as PostScript),
interpreting PDF is still significantly slower than using a resident
rendering session when font setup comes into play.  So for things like
refreshing display and processing thousands of files into bitmaps (which
we actually do when generating the web-based documentation), direct
Cairo rendering would be a big boon.

-- 
David Kastrup