Just a bit of news to report.
The full set of tests for citeproc-js are now taking 30 seconds to run
on my little Atom netbook, and I started casting around for ways of
running this in spidermonkey, and as a persistent process that doesn’t
require re-“compiling” for every transaction. This morning I hit pay
dirt with python-spidermonkey:
http://code.google.com/p/python-spidermonkey/
This bridge provides enough support to load the processor code to a
Python object, feed input to it, and get back rendered strings. A few
details need to be worked out, but it should be a fairly short step
from here to a simple demo web service such as panyansan has
requested.
This little discovery has cheered me up a bit, so I thought I would
share it with the group.
Frank