Wikinews:Water cooler/technical/archives/2011/January


Facebook

Does anyone know why our Facebook page has stopped updating? Have I missed a previous discussion on this? wackywace 12:40, 23 December 2010 (UTC)[reply]

I blame ShakataGaNai! :P Bawolff 12:48, 23 December 2010 (UTC)[reply]
What happened? I would be happy to operate the account, unless someone else want to. wackywace 12:50, 23 December 2010 (UTC)[reply]
I've been wondering for quite a while also, if auto-update cannot work I'd gladly offer myself to update it along with Wackywace. --Bobby122 Contact Me (C) 14:07, 23 December 2010 (UTC)[reply]
I attempted two or three weeks ago to set up a different Wikinews page with auto-updating... but I finally gave it up after hours of wasted time (especially since the RSS feed was down at the time). If somebody wants to operate that page, I'd be happy to give the admin status there. Diego Grez return fire 15:41, 23 December 2010 (UTC)[reply]
I'd be happy to, and two heads are better than one, so if Bobby wants to do it too that'd be good. The only problem I can see is that the old account has 2,000 fans—the new one has two—telling all 2,000 of them to become fans of the new one would be impossible. Does no-one actually have the password to the old one? Is it lost forever? wackywace 15:59, 23 December 2010 (UTC)[reply]
SGN has it. Blood Red Sandman (Talk) (Contribs) 16:00, 23 December 2010 (UTC)[reply]
Does he still contribute here? If not, does anyone have contact details? wackywace 16:07, 23 December 2010 (UTC)[reply]

I've been bold and gone ahead and created a new one. Upon reconsideration, I think a new page is the best option—it will be odd if the old one just starts up again like nothing happens, and it will be a fresh start and a chance to see how Wikinews is doing. Bobby, if you e-mail me (click the link in the box that says e-mail user here) your name so I can find you on Facebook (I presume you are on there) and I'll try and find how to make you an admin. Any other trusted contributors are, of course, welcome to help out too. wackywace 18:14, 23 December 2010 (UTC)[reply]

Shakataganai still contributes here (and is usually on irc, especially on #wikimedia-commons) channel. I suggest just getting the password for the current one from him. No point starting ten different pages. Bawolff 06:09, 24 December 2010 (UTC)[reply]
bastique aka Cary Bass has access to the page. I was an administrator on the Wikinews page, but revoked my own rights as it interfered with my actually FB account. He can set someone up with admin there. DragonFire1024 (Talk to the Dragon) 06:15, 24 December 2010 (UTC)[reply]
In regards to Wackywace's comment, what type of rss feed do you need that currently does not exist? Bawolff 06:20, 24 December 2010 (UTC)[reply]

  • I'll chase Shaka for sundry account details, but I do have a suggestion; all needed code goes into the public domain, and everything gets driven via identi.ca. That provided with updates through running a few cron jobs on wikinewsie.org.
I'm getting tired of seeing these things periodically break down; the amount of duct tape, bailing twine, and spit holding together Wikinews' social networking is shameful. --Brian McNeil / talk 08:59, 24 December 2010 (UTC)[reply]

You guys realize this isn't an "account", it's a page. Give me a name and I can make them an admin. --ShakataGaNai ^_^ 20:49, 25 December 2010 (UTC)[reply]

Let me be clearer of the process to "fix" this.
  1. Kill Alt 1 and Alt 2
  2. Friend me, make sure to include (in the msg) who you are on wiki (in case it isn't obvious)
  3. I will make you an admin on the main Facebook page.
  4. Someone figures out how to make it update again
  5. Everyone goes home happy
Easy enough? --ShakataGaNai ^_^ 04:42, 26 December 2010 (UTC)[reply]
  • First problem is, how many of the grumblers have ever cut any code? This should not be running off the RSS feed. Since I've managed to hack wikinewsie.org into on-demand scraping the latest few headlines, I can probably resolve scraping them every 5 minutes, checking for diffs (i.e. published since last scrape), and pushing to Farcebook. The RSS feed is a <bleep> <bang> <$£&#> of excrement. --Brian McNeil / talk 09:18, 26 December 2010 (UTC)[reply]
Why re-invent the wheel. The rss feed parses T:LN, and outputs it in a sensible format. We'd need a screen scraper of some sort to power this, might as well use the one we've already got. Its also the part of this system that breaks the least. Bawolff 10:48, 26 December 2010 (UTC)[reply]
  • Why use a <spit> 'screen scraper'? The API should be in-use. Yes, the correct solution is the Google News feed extension, but we all know how tricky it can be getting things like that resolved. I'm appalled at what Shaka has laid out as the data flow; just how many points-of-failure in there? --Brian McNeil / talk 12:05, 26 December 2010 (UTC)[reply]
You want to get the latest published articles (in cat published, not in a bunch of other categories), that are sighted. Unless you want to do it in multiple api hits, you're going to have to screen scrape since such functionality (afaik) is not in api since its only provided by DPL. But in any case, the system is far from perfect. However even if the facebook feeder script was getting its data in some nice clean way, it'd still be broken, because as far as i understand its the post to facebook step that is broken, not the get latest news step. Bawolff 12:21, 26 December 2010 (UTC)[reply]
  • The post to FB step is borked because FB hacked the hell out their API; you can no longer do it in an automated manner without actually creating an App. That ties to into an arms race of following every whim of theirs (i.e. enforced development at their whim). However, if we go thru identica - using the WNdent account - they're left doing all the catchup, and I'd expect reasonable lead time on interface changes. I can probably sort a fair bit of that. --Brian McNeil / talk 12:58, 26 December 2010 (UTC)[reply]

Facebook et-al, drive from WNdent

Notes:

  1. Statusnet app for FB update: http://apps.facebook.com/identica/
  2. Auto-post from identica to Twitter: http://status.net/wiki/Twitter

This should be the basics to sort out some of this. Can't get at the FB link whilst in the orifice. Already have some of the requisite code to scrape from a DPL on wikinewsie.org. All I require for this is:

  1. TIME
  2. Username and password for the twitter account

As Shaka has reminded me I have Admin access on the FB account, I can resolve this. Amgine is holding the WNdent account on identica, so I'll be able to help arrange that one being resolved.

What I'd propose doing is a 'scrape' of the latest news DPL every 5 minutes. Any newly published article would then be pushed out.

In relation to this, is anyone aware of the current status of the Google News extension? I believe this was kicked back into play post-SoC. --Brian McNeil / talk 13:53, 26 December 2010 (UTC)[reply]

post summer of code? GSoC did not affect its status as far as I'm aware. Hexmode has indicated he might work on it. At the moment there are still issues with the extension that need addressing (they're documented on CR. I can dig up the link if you want.) It is probably not getting enabled in the very near future, but may be at some point. Bawolff 14:43, 26 December 2010 (UTC)[reply]
Brian, what is it you're trying to solve/do? Facebook part I get and you've got Admin for that. Twitter I'm confused about. Twitter is already being updated by Identica. Also, I (not Amgine) own the Identi.ca account (Along with Twitter, Along with Google Buzz). --ShakataGaNai ^_^ 07:04, 3 January 2011 (UTC)[reply]

Template:Editintro notcurrent

When editing an older article, this template shows up on the top of the page. This is not a problem, but its interwiki link also appears. I have checked on the template, and the interwiki section is already nested in a <noinclude>. I have no idea how to fix it. - Xbspiro (talk) 15:31, 16 January 2011 (UTC)[reply]

Noinclude doesn't work in editintro templates. You need to use {{#ifeq}} magic. Should be fixed now. Bawolff 03:06, 17 January 2011 (UTC)[reply]
I fixed the issue in rev:80434 (assuming i don't get reverted). However that fix probably won't go live for something like 9 months (the 1.18 release). In the mean time, we need to use the #ifeq hack. Bawolff 03:36, 17 January 2011 (UTC)[reply]
Thumbs up. - Xbspiro (talk) 13:15, 19 January 2011 (UTC)[reply]

category redirects.

Redirects in category namespace are currently discouraged, since they appear as a bluelink at the bottom of the page, and thus result in divided categories. I added css to make them appear as red links at the bottom of the page, to help address this issue (may have to do hard refresh to see). Bawolff

Yeah, I wish i knew mediawiki output classes for those redirects earlier. The info's been there are along, and we've just never used it. There's a bug for adding it to standard css (bugzilla:5346), Bawolff 00:12, 20 January 2011 (UTC)[reply]