This is an experimental backend that put the LibHaru PDF module to work in the PDF tool-chains. Note that LibHaru doesn't read PDFs, but what's there looks good.
This is a bottom-up integration, which is centered on representing and serializing PDF objects. The intent is to map Dicts, Arrays and primary objects to LibHaru objects.
Hopefully access may ber somewhat quicker and with a lower memory footprint.
Serialization, has to be quicker, we can directly output dictionarys and arrays to memory or files.
I'm not so interested in the high level stuff: Pages, Catalogs, etc. That can be done in Perl.
There's some other areas of the toolkit that may be of interest such as encodings, fonts, xrefs.