Spay or neuter your jQuery libraries before they multiply!

If any MediaWiki hacker folks are looking for an easy introduction to working with ResourceLoader, one way might be to help with bug 26306 “Cleanup extensions that include jquery (and common plugins) themselves”.

A number of extensions over the years have pulled in their own copies of jQuery, or particular jQuery or jQuery.ui plugins which are now present in MediaWiki core and available through ResourceLoader.

The extra copies — which are usually horribly out of date! — can generally now be removed.

Here are commits I made on a couple extensions:

  • r87845 OpenID: needed conversion to RL, used the opportunity to clean up other minification/spriting
  • r87847 Collection: most of the old modules had already been switched to RL but the old files were still there. Cleaned up and swapped one more plugin for the core copy

Don’t be afraid to check out the migration guide for extension developers and related doc pages on, or bug your fellows — anything that’s not clear now is something that should be documented better, so don’t be shy!

— brion

Weekly whinge: offline feed reader for Android?

Google Reader for android is great… if you’re always online. If you have a subway commute, fly a lot, or otherwise like to catch up on your blogoverse when the rest of the world isn’t talking to your phone, you’ll quickly discover that the app’s offline sync feature is poor enough to be more frustrating than not having it at all.

They seem to have done a lot of things right; starring or marking things read works just fine while offline and queues the server updates transparently for later.  And when an article is available offline it fetches it cleanly and quickly.

There are basically three problems:

First, there’s no way to tell what feeds will actually be pre-downloaded. It claims to fetch your “most read feeds” but it gives no indication which those are. After several weeks of use I’m still getting confused by what is or isn’t being fetched.

The second problem is that the list displays do not distinguish between new feed entries *that have been reported to exist by the server* and those that have already been downloaded. You can see a clear count of entries and be faced with a “no network; retry” screen. Worse, if the connection is intermittent it’ll try to connect for a few seconds before telling you there’s nothing to see.

And to add insult to injury, images are not prefetched, consigning subway warriors to a pure-text life devoid of photos, screenshots, charts, icons, diagrams, maps, etc.

I could probably whip something up that does what I want, but polishing it won’t be trivial and I’m sure there are other tools out there that would suit my needs; anybody know a good offline-friendly feed reader of Android, preferably with Google Reader sync or import?