All The Tropes Wiki Project, Part V

[GethN7]

The Trope Workshop still looks screwy.

 

The 'create a new article" bit with the InputBox extension is broken due to a known bug with InputBox, everythign else is fine.

You can always make a new workshop page manually.

[Bob Schroeck]

... Just tried my bot again. Just got 502 again, at the same point in the code -- retrieving a pagelist. You sure I'm not breaking the wiki? Maybe I'm trying to pull too much too fast?
-- Bob
---------
Then the horns kicked in...
...and my shoes began to squeak.

[GethN7]

... Just tried my bot again. Just got 502 again, at the same point in the code -- retrieving a pagelist. You sure I'm not breaking the wiki? Maybe I'm trying to pull too much too fast?

 

It's coincidence. You turned your bot on just around the time some LC issues were reported on IRC.

Turns out they are sorting out some DB issues that cause periodic PHP crapout, hence all the downtime.

On that note, ATT's database has a few issues that are being resolved: All text is confirmed fine, though a few orphaned revisions were discovered (this is mostly harmless). As for images, those need the image table rebuilt and the images reimported, our earlier DB issues there caused that to go wonky, but that part is behind us, we just need the reimport of the images done.

We ARE getting better,. and the Miraheze guys are debugging things, that I can verify.

[Bob Schroeck]

Trying again, starting at 10:05 PM EDT... and I got the 502 error again at the same point. I must be causing it.
-- Bob
---------
Then the horns kicked in...
...and my shoes began to squeak.

[GethN7]

Trying again, starting at 10:05 PM EDT... and I got the 502 error again at the same point. I must be causing it.

 

It's possible that is triggering it.

Since bots have to query the MediaWiki API, which fetches items from the database, this could be causes the DB errors to assert themselves and result in 502s.

If this doesn't clear up soon, I'll be making a more pressing request to get this fixed PDQ.

[GethN7]

Bob, Drew got back to me on that infringing page. He has offered to either review the allegations and simply remove the page or, and I quote his exact message here:

"Hi Geth N 7 - I'll just change the CC license for this page to match yours. Good enough for you?"

What would you find acceptable?

[Bob Schroeck]

Oh, wow. That's actually a difficult choice. Strike a minor blow against TVT, or insert a wedge in its copyright stance? If we get them to set a precedent for one page, even if doing so is dubiously-legal ass-covering... (Can they actually change the license for just one page? Or does that have repercussions or limitations?)
-- Bob
---------
Then the horns kicked in...
...and my shoes began to squeak.

[GethN7]

Oh, wow. That's actually a difficult choice. Strike a minor blow against TVT, or insert a wedge in its copyright stance? If we get them to set a precedent for one page, even if doing so is dubiously-legal ass-covering... (Can they actually change the license for just one page? Or does that have repercussions or limitations?)

 

They can license certain aspects of their site to a certain copyright license, just as we can, though much like us, they have to clearly define on those pages which license applies.

Should you go that route, I would insist that (A) they do that and (B) they clearly link back to us, since that is required for CC BY SA attribution.

[Bob Schroeck]

Let's go for that. In case my original post on the subject didn't make it clear () I don't mind my work being echoed there.
-- Bob
---------
Then the horns kicked in...
...and my shoes began to squeak.

[Bob Schroeck]

Oh, by the way, guess what I tried at 10:41 PM. Guess what happened at 10:42 PM.

I'm going to give AutoWikiBrowser a try. Maybe it's DotNetWikiBot's attempt to batch-process pages that's causing this.
-- Bob
---------
Then the horns kicked in...
...and my shoes began to squeak.

[GethN7]

Oh, by the way, guess what I tried at 10:41 PM. Guess what happened at 10:42 PM.

I'm going to give AutoWikiBrowser a try. Maybe it's DotNetWikiBot's attempt to batch-process pages that's causing this.

 

If it's trying to handle multiple pages at once, yeah, that could cause a crapout since it would be hammering the API with multiple read/write requests.

Usually, it's safest to have it have like a 3-5 second interval between each page edit (bare minimum) to avoid hammering the API too hard. AWB comes with built in options to add a breathing period inbetween page edits to avoid this very issue.

[robkelk]

Just started getting 502s across Miraheze...

(At least I recovered four pages from various caches for the freebie wiki before that started.)
--
Rob Kelk
"Governments have no right to question the loyalty of those who oppose
them. Adversaries remain citizens of the same state, common subjects of
the same sovereign, servants of the same law."

- Michael Ignatieff, addressing Stanford University in 2012

[GethN7]

Just started getting 502s across Miraheze...

(At least I recovered four pages from carious caches for the freebie wiki before that started.)

 

They finally figured out the source: HHVM keeps dying on them at occasional intervals, and certain API calls cause it to peak and die for some reason. They are working on localizing why.

[Labster]

"Hi Geth N 7 - I'll just change the CC license for this page to match yours. Good enough for you?

Yeah, TVT is allowed to use our stuff if they use the same license.  If they broke the agreement by posting under another license, then TVT's 3.0 license was terminated, but since we dual license under 4.0, he can take remedial action under that one.  Therefore, he can go ahead and use our content without permission with the correct copyright (and a backlink for attribution clause).

Just started getting 502s across Miraheze...

Meet the new wiki farm, same as the old wiki farm...
Also I'm taking steps to move the wiki to http://www.allthetropes.org/  You know, so we don't lose all of our SEO in the future.  Hopefully Orain can throw up some redirects in the end.

rrrgghh...  That worries me.  Looking at Drew's credentials, I can tell
that this guy is a suit - full stop.  I don't think that means good
things for TVT, because this guy's end goal is to monetize shit.  And
somehow this rubs me wrong even worse than Fast Eddie's Stalinism.

Yeah, monetization is definitely a goal, but so is improving the site.  His credentials don't tell the whole story -- in person he told me that he hated patent lawyering, so he joined his friend's startup to get away from all of that.  If I were running it, I'd leave the wiki pretty much alone, and use it to funnel people into the startup, Share TV.  And maybe use Share TV's ad network to supplement TVT's income.  But as always money forces content tradeoffs -- if you're not paying, then you're the product.  This is why we chose the nonprofit model.

Basically, we'd have three branches that could share content, you can stick with the one you prefer.

Of course, this is not a done deal. Orain has to prove to me they can
restore us to our antebellum condition first or as close as possible. If
they can't, this is off the table.

Yeah, to me, Orain is off the table.  It's not the crisis, it's the crisis response.  Apparently Dusti isn't around because he's working 12 hour days.  (First of all that will probably be seen as a human rights abuse a century from now.)  But the key person cannot be unavailable for so long.  The other key person, kudu, has had only sparse availability.  And, this has happened before, and John Lewis kept things together without the key people.  This time John wasn't here because of the falling out, and now no one is cleaning up the fall out.
-- ∇×V

[GethN7]

Well, here's proof at least this wiki farm is working on the 502 issue:

http://github.com/miraheze/mw-config/issues/162

[Bob Schroeck]

And while they bang away at that, I'm running into problems with AutoWikiBrowser. When I try to set the site URL in the preferences, I get an error dialog reading:

Error connecting to wiki

An error occured while connecting to the server or loading project information from it. Please make sure that your internet connection works and such combination of project/language exist.

Enter the URL in the format "en.wikipedia.org/w/" (including path where index.php and api.php reside).

Error description: Root element is missing.

Now I've tried every obvious choice -- ending the path with "en/" and with "wiki/" among others -- but I keep getting this. What's the right thing to do?
-- Bob
---------
Then the horns kicked in...
...and my shoes began to squeak.

[GethN7]

And while they bang away at that, I'm running into problems with AutoWikiBrowser. When I try to set the site URL in the preferences, I get an error dialog reading:

Error connecting to wiki

An error occured while connecting to the server or loading project information from it. Please make sure that your internet connection works and such combination of project/language exist.

Enter the URL in the format "en.wikipedia.org/w/" (including path where index.php and api.php reside).

Error description: Root element is missing.

Now I've tried every obvious choice -- ending the path with "en/" and with "wiki/" among others -- but I keep getting this. What's the right thing to do?

Try this:

allthetropes.miraheze.org/w/

[GethN7]

FYI everyone, getting the Trope Workshop fixed up soon, Miraheze lacked Orain's config settings for it.

[Bob Schroeck]

Try this:

allthetropes.miraheze.org/w/

Now getting an error about not being able to find Newtonsoft.Json.dll of the right version. So far the fixes I've found online have done nothing.
-- Bob
---------
Then the horns kicked in...
...and my shoes began to squeak.

[GethN7]

Try this:

allthetropes.miraheze.org/w/

Now getting an error about not being able to find Newtonsoft.Json.dll of the right version. So far the fixes I've found online have done nothing.

 

Make sure all the dependencies for whatever system you're using are up to date, sounds like a .NET issue.

[GethN7]

Bad news, all:

http://github.com/miraheze/mw-config/issues/172

The Trope Workshop page revisions are "orphaned", which means they are in the database, they just aren't linked to the proper pages, so they exist, but only as "phantoms".

The issue above is for them to fish through the database to find those revisions and restore them to the proper pages.

Update: Already fixed.

[Bob Schroeck]

Database error

A database query error has occurred. This may indicate a bug in the software.

Function: SearchMySQL:earchInternal
Error: 145 Table './allthetropeswiki/searchindex' is marked as crashed and should be repaired (185.52.1.77)

-- Bob
---------
Then the horns kicked in...
...and my shoes began to squeak.

[Bob Schroeck]

Why is it that when I add a page to my watchlist by clicking the star icon at the top of the page, some pages go right into the list, and others require that I confirm I want them? And I presume the feature where you didn't actually leave the page when you added it to your watchlist has yet to be reimplemented?

Meanwhile, may I just say Argh! The rather large and detailed article I did up for Lord Buckley was one of those lost. Gah. I don't look forward to recreating that any time soon. However, I did manage to rescue Reciprocal Real Person Cameo and the Trope Workshop's Devil's Dictionary from the Google cache. When I get a chance I'll convert them back to wiki markup.
-- Bob
---------
Then the horns kicked in...
...and my shoes began to squeak.

[GethN7]

I was warned the search might crap out like that. John said he'd try to give us CirrusSearch again, going to check into the status of that.

Also Bob, update from Drew, the page you brought up as a copyvio has been cut.

[Bob Schroeck]

Interesting. I wonder why they chose to do that when I was happy with the license change. Because of the precedent, maybe?
-- Bob
---------
Then the horns kicked in...
...and my shoes began to squeak.