I suspect that many will say this is long overdue: here has been a
steady trickle of firmly worded complaints about the awkwardness of
interacting with the citeproc-js sources.
Years ago, I attempted a conversion of the citeproc-js sources to git.
At that time, the conversion failed due to an unnamed head in the hg
revision tree. It’s actually very hard to remove one of these, and I
gave up. Apologies now for the years of inconvenience that followed.
The issue has come up again on my own desk, with a push to clean up
the documentation picture for citeproc-js, CSL-M, and Juris-M. The
obvious path forward is to lodge things in ReadTheDocs, and as I’m
already familiar with web-hooks in GitHub, moving citeproc-js over
there will save.
I’ve just taken a shot at conversion with fast-export
(https://github.com/frej/fast-export), and it is able to step past the
unnamed-head error with a --force option. The code that comes through
in git is correct: the test suite clears.
So moving to GitHub is possible. I’m thinking to do the following:
- Add a REPO-MIGRATION note to the citeproc-js sources on BitBucket.
- Push citeproc-js to the Juris-M space on GitHub.
- Set up the processor manual on ReadTheDocs.
- Remove the source from BitBucket, leaving behind the REPO-MIGRATION note.
Many on this list have a lot more experience with this sort of thing
than I do. Any suggestions on how to avoid general chaos would be most
welcome.
Frank