When we refatored the core, I added a plugin system. It has turned out that it was not really used because it really is nearly identical to the extension idea. It even horribly confusing (huh, what's the diff between the two?). I have completely removed the plugin system.
Removing plugin module and its tests.
When we refactored the core we created a plugin system, but we
don't use it for anything and it is completely redundant (we
already have extensions).
Cleaning up extensions that used plugins.
Yes, and I think that with the refactored magics that can cleanly hold isolated state, the last bit of justification for these is gone. The code looks clean, let me run test_pr for sanity.
I think we should give it a day or two for others to pitch in, in case someone really disagrees.
The only comment I'd make is that now is the time to add a note to the docs, in the what's new document, backwards-incompatible changes, mentioning this big removal. Otherwise we'll forget later.
Test results for commit 1e6eb00 merged into master (d3f6cea)
Not available for testing: python2.6
Makes sense to me. Ping the list, in case anyone ever actually wrote a plugin?
Did the list get pinged about this? I don't see anything with 'plugin' in the subject in the archives for August.
LIst has still not been pingged I think. I would prefere SO which speek better english than me, and who knows the pluggin system a little more to ping the list, but I can do it.
Let me know.
I have now pinged both lists. For reference, here's the message: http://mail.scipy.org/pipermail/ipython-user/2012-September/011255.html
So far (3 days), no-one has objected.
Thanks for pinging the lists, I don't think there is any objections so I am going to merge.