The citeproc-js sources have been moved, from Sourceforge (svn) to a
new home at BitBucket (Mercurial/hg). The new address for the sources
and issue tracker is:
BitBucket is based on the Mercurial (hg) version control system, so to
follow the sources from here out you will need to have hg installed
and clone the archive. The command line incantation for making a
clone is shown at the URL above.
This is product of some off-list discussion between Frank and I, and
my general sense that as we move towards finishing CSL 1.0 and getting
the test suite in place, we can and should strip down the xbib project
repository and refocus.
It seems to me BitBucket and GitHub are both better options to host
implementations, because they have both have similar, really nice,
interfaces for mercurial and git (respectively) repositories and
At the same time, while SF has added some stuff recently, I don’t
think it presents a very good package for us. For example, while I
have git and mercurial support switched on, it seems SF only allows
one repo per project. That just sucks (unlike CVS and SVN, you really
need to put subprojects into separate repos)…
We also have a little problem in that citeproc-rb is hosted here, but
someone has forked it on GitHub. Perhaps it’d be better to just throw
out the revision history of that subproject, convert it to git, and
move it over to GitHub so that all of us could work on that?
I’ve already put the start of a citeproc-py rewrite on GitHub, but
have the problem that I can’t really use Johan’s code in a non-GPL
project until he changes the license.
So we have citeproc-js and citeproc-hs looking in good shape and now
with nice homes. I’d like to see if we could do that same with
citeproc-rb and citeproc-py.
That way, xbib really becomes more csl, and we just host the schemas,
locales files, documentation, and test suite.
I have been working with BitBucket for all of one day, but it seems
very smooth and zippy. If you are working with or following the
processor and run into an issue or have a test, feel free to make use
of the tracker, or drop me a line if you have changes you’d like to
have pulled into the archive.
Development in the SVN archive will stop at this point. I’m not quite
sure what to do with the sources there, but as the processor is still
unfinished, my intuition is to just delete the sources and replace
them with a note pointing at the new URL. I’ll let the issue rest for
about a week, and go ahead with that if no one has any objections.
If we do a similar thing with the rest, we might just move a lot of
legacy stuff to an attic directory?
As for the state of the project itself, I do think we’re getting close
to the end of Round 1. I’m hesitant to offer a forecast, since I
think I’ve underestimated the difficulty and the complexity of just
about every major CSL feature I’ve tackled to date, but I do believe
the largest dragons have been put to rest. Within a month or so I
hope to be running tests against production styles.