Doom Wiki

Xeriphas1994 said:

Are user preferences preserved in the fork?  E.g., if I've been receiving automated emails about articles on my watchlist, will I start getting them for the new wiki also?

mysql> desc watchlist;
+--------------------------+------------------+------+-----+---------+-------+
| Field                    | Type             | Null | Key | Default | Extra |
+--------------------------+------------------+------+-----+---------+-------+
| wl_user                  | int(10) unsigned | NO   | PRI | NULL    |       | 
| wl_namespace             | int(11)          | NO   | PRI | 0       |       | 
| wl_title                 | varbinary(255)   | NO   | PRI |         |       | 
| wl_notificationtimestamp | varbinary(14)    | YES  |     | NULL    |       | 
+--------------------------+------------------+------+-----+---------+-------+
4 rows in set (0.00 sec)

mysql> select * from watchlist;
Empty set (0.00 sec)

That is a big no, you will have to reset your watchlist. The MediaWiki API [1] provides access to watchlists, but apparently MediaWikiDumper doesn't scrape it.

[1]http://doom.wikia.com/api.php

spicyjack said:

That is a big no, you will have to reset your watchlist. The MediaWiki API provides access to watchlists, but apparently MediaWikiDumper doesn't scrape it.

OK, good to know, thanks.  I guess this might be a security precaution: otherwise, a vandal could take the dumped output and compile a list of all unwatched pages, then auto-post crap on them (which IIRC has happened on Wikipedia).

My own watchlist is no big deal, but I was thinking about inactive editors who receive notifications of talk page comments.  Now they have to take additional action to get the new ones.

Speaking of wikis... Anybody knows a way to get a list of changes older than 30 days in MediaWiki? No amount of URL editing seems to work.

I ask because many Doom-related wikis have suffered attacks from spammers recently, and those that aren't patrolled by admins able to delete spam are suffering from it. And if someone were to decide to clean up, being able to only view unwanted edits and additions from the last month would be quite a hindrance.

I hate the new design.

Mainly because of all the ads, especially moving ones. It destroys my virtual memory, norton system works pops up every few minutes saying that virtual memory low and doom wiki is the cause.

And all the pop ups as well, piss me off.

Everything takes forever to load, too.

You should see how it looks on PS3 web browser, though....

Thanks to spicyjack's help the Doom Wiki has been imported onto doomwiki.org/.net/.info. Please cease editing the Wikia wiki at this point if you intend any of your subsequent edits to appear on the new wiki when it comes online, as we are unable to port any additional edits across (the script will not update for some reason, it only wants to regrab the entire wiki every time and we can't do that again).

There is an absolute ton of cleanup which needs to be done, but we hope to open it to the public as soon as humanly possible.

Gez said:

Speaking of wikis... Anybody knows a way to get a list of changes older than 30 days in MediaWiki? No amount of URL editing seems to work.

The database will tell you.

Probably easier to find posts by spammer lusers and then work through that user's contribution special page.

mysql> select rc_timestamp, rc_user, rc_user_text, rc_title from recentchanges where rc_user = 422 limit 5;
+----------------+---------+--------------+---------------------------------+
| rc_timestamp   | rc_user | rc_user_text | rc_title                        |
+----------------+---------+--------------+---------------------------------+
| 20101124181738 |     422 | Gez          | Doom_Resurrection_(WAD)         | 
| 20101124090602 |     422 | Gez          | MAP28:_The_Absolution_(Doom_64) | 
| 20101123180957 |     422 | Gez          | Central_Processing              | 
| 20101123180443 |     422 | Gez          | GZDoom                          | 
| 20101123180251 |     422 | Gez          | Urban_Brawl                     | 
+----------------+---------+--------------+---------------------------------+
5 rows in set (0.00 sec)

@yellowmadness54
Adblock plus. Use it. All ad woes gone away.

Quasar said:

Thanks to spicyjack's help the Doom Wiki has been imported onto doomwiki.org/.net/.info.

That's awesome! A sincere Thank You to everyone involved!

I'm uncertain how we rank-and-file wiki editors can convey our gratitude to the new backend people, in a way that will actually mean something.  Just yelling "awesome!" in a forum, or on a discussion page of the new site, doesn't seem like enough.

spicyjack said:

The database will tell you.

With respect, this is an incomplete answer -- some wiki admins are not programmers.  Aren't there config options for stuff like that, though?  "$wgMaxDBRows" or whatever?

In the long term, will independent hosting allow the Doom Wiki to create script-assisted vandalism tools (like Twinkle/Huggle/etc. on wikipedia), so that the non-programming editors can clean up almost as quickly?

Xeriphas1994 said:

With respect, this is an incomplete answer -- some wiki admins are not programmers.

More to the point, some wiki admins do not have direct SQL access to the database. Or HTTP access to modify PHP files, either.

Gez said:

More to the point, some wiki admins do not have direct SQL access to the database. Or HTTP access to modify PHP files, either.

Well, IIRC the second thing can be changed if Wikia isn't moderating, and in theory we could even set up a more direct interface for mass edits (like Wikipedia's toolserver).  But that stuff is for later -- I didn't want to sidetrack the goal of getting the new site "open for business".

I do agree with Gez that anti-vandalism measures will be needed sooner rather than later (as is the case on all wikis nowadays, everywhere).  My point was that if longer recent changes lists can be obtained by changing *one switch* in the config file, that seems like a reasonable request.

Xeriphas1994 said:

In the long term, will independent hosting allow the Doom Wiki to create script-assisted vandalism tools (like Twinkle/Huggle/etc. on wikipedia), so that the non-programming editors can clean up almost as quickly?

Allow, sure, but who is going to write them? I'm not a php OR MySQL programmer, so don't count on me doing it ;)

I hope that no one is *counting on* anyone to do anything, other than what they have already volunteered, which sounds time-consuming and unpleasant yet is being done.

I was merely suggesting that, in addition to the new technical obligations of independent wiki hosting (which should have their own thread on the wiki I think), we also have new technical opportunities.

Xeriphas1994 said:

I hope that no one is *counting on* anyone to do anything, other than what they have already volunteered, which sounds time-consuming and unpleasant yet is being done.

I was merely suggesting that, in addition to the new technical obligations of independent wiki hosting (which should have their own thread on the wiki I think), we also have new technical opportunities.

Right. What I'm saying is, though, we may not be able to find anyone with the requisite tech skills *and* the desire to do that kind of stuff. It *would* be great, of course.

Gez said:

Speaking of wikis... Anybody knows a way to get a list of changes older than 30 days in MediaWiki? No amount of URL editing seems to work.

I ask because many Doom-related wikis have suffered attacks from spammers recently, and those that aren't patrolled by admins able to delete spam are suffering from it. And if someone were to decide to clean up, being able to only view unwanted edits and additions from the last month would be quite a hindrance.

Eternity Wiki to 09Apr2010:
http://eternity.youfailit.net/index.php?title=Special:RecentChanges&days=360&limit=1000

ZDoom Wiki to 29Nov2010:
http://zdoom.org/w/index.php?title=Special:RecentChanges&limit=1000&days=360

Vavoom Wiki to 03Feb2011:
http://www.vavoom-engine.com/wiki/index.php?title=Special:Recentchanges&limit=1000&days=360

Wikia's Doom Wiki to 16Jan2011:
http://doom.wikia.com/index.php?title=Special:RecentChanges&days=360&limit=1000

With my copy of the Doom Wiki and the mostly default install of MediaWiki, I can go back to 10Oct2010.

So I'm guessing this is a setting in MediaWiki that individual Wiki admins can change. If you want something different, I suggest talking with the admins of the above wikis and asking.

Some snooping in include/DefaultSettings.php gives me:

/**
 * Recentchanges items are periodically purged; entries older than this many
 * seconds will go.
 * Default: 13 weeks = about three months
 */
$wgRCMaxAge = 13 * 7 * 24 * 3600;

This is the default setting. If you want more than that, you'll need to crack the database.

Ha, there's also an option to relay recent changes;

/**
 * Send recent changes updates via UDP. The updates will be formatted for IRC.
 * Set this to the IP address of the receiver.
 */
$wgRC2UDPAddress = false;

Anyone want to try getting CIA to report wiki changes, or would that make the bot far too noisy in IRC?

It was probably premature to announce this. We've found that for some reason a lot of critical pages are red-linked. They either weren't downloaded or the database is FUBAR. Examples of what I mean by critical would be "ZDoom", for example.

The page count is +11000 when I do a count(*) in mysql but the Special:Statistics page lists 0 articles. This makes it extremely difficult to even figure out how many articles are apparently missing.

And in the meanwhile, I have neither heard from nor seen anything of Manc for 4 or 5 days now, and we're waiting on him because he hasn't given the doomwiki account enough permissions to proceed with the remainder of the process. Even the maintenance.php script cannot be run (and it clearly needs to be run) because the php script memory limit is set far too low.

It's understandable that unforseen circumstances can cause delays, so no worries there. We're not expecting an exact timeframe and we all appreciate the work you're doing, so don't feel under any pressure if things aren't quite going to plan.

DooMAD said:

It's understandable that unforseen circumstances can cause delays, so no worries there. We're not expecting an exact timeframe and we all appreciate the work you're doing, so don't feel under any pressure if things aren't quite going to plan.

Yeah ;) A clean import process it has certainly not been. Anyway Manc and I are back in touch and hope to iron out some stuff tonight (or try anyway ._.)

My HOPE is that running maintenance.php successfully will magically heal the database and the red links will go away. That is pretty optimistic though. In a bad-case scenario we're going to have to do the rip again. In the worst-case scenario, the rip isn't ever going to work right and we're going to have to import from the wikia db dumps, which are missing a ton of stuff we already do have and I'd rather not lose.

What we may really end up *needing* is something that could only grab the stuff we don't have already out of the db dump. I don't have a clue if that's possible or not, unfortunately, so hopefully it won't come down to that.

spicyjack said:

So I'm guessing this is a setting in MediaWiki that individual Wiki admins can change . . .

/**
 * Recentchanges items are periodically purged; entries older than this many
 * seconds will go.
 * Default: 13 weeks = about three months
 */
$wgRCMaxAge = 13 * 7 * 24 * 3600;

To clarify the code comment: diffs older than the default age may be corrupted or missing, because the underlying table only keeps "pointers" for that number of entries.  Therefore, simply raising the limit is not guaranteed to turn up every change; you also need to rebuild the table somehow from the individual page histories.  Or scrape it, I suppose.

(Sorry for not mentioning this immediately.  Wikia's developers explained these limits on the listserv years ago, but I didn't realize that the hard display limit and the table size limit were now the same.  Yuck.)

For the record, no one contacted me about needing specific help. I installed perl modules and a couple of other minor things that were requested as soon as I could do them, but beyond that no one has come to me with these issues. I was not on irc on monday (valentine's day, real life and all). I was on IRC on tuesday. I was not wed. or thursday. Not to mention I was on all weekend as my comments in various channels can attest to. Especially considering some of the real import work was only done a couple of days ago. There are a variety of ways to contact me. Pretending I'm glued to all channels is not one of them.

We need a way to additively import to the database from a wikia dump or this is never going to be finished. If such a thing doesn't exist, then the database cannot be properly imported, period.

So if you know anything about php, MySQL, and wikis, please help. Otherwise we're going to have to settle for a broken wiki with a shitload of missing articles.

Even if that anything is merely "a few years ago I've customized a bit a Wakka wiki engine to let it accept MediaWiki-style titles with spaces and accented letters instead of just CamelCasedWikiWords?"

If it really comes down to it, I'm sure it's possible to write a scraper that just grabs the raw MediaWiki markup from the each page's edit page, and then a "poster" (?) to create a new page and post the markup in that page's edit page. I don't know what to do about images and such though.

Thanks again to spicyjack, we were able to get a clean dump. The database may not be 100% perfect (I've noticed some falsely-red links that still take you to the article if you click them) but we haven't found anything missing so far.

Cleanup work will be commencing very soon.

Quasar said:

Cleanup work will be commencing very soon.

Impressive news!  Many thanks to the brave souls doing the actual config work.

For those not following on IRC, "cleanup" refers specifically to recruiting interested wiki admins (who presumably have been around a while and know what everything is *supposed* to look like), for help in removing the Wikia leftovers and making sure links/editing work normally.  Um, so are any additional wiki admins lurking in this thread?  :D

Progress is happening. I have expunged most of the references and links to Wikia in the article, MediaWiki, Help, and Doom Wiki namespaces. Manc has installed the Monaco skin. However, I am bothered right now by the fact that it doesn't display the copyright notice anywhere.

I took a look into the php and I'm totally lost. I found the code that generates the footer and customizing it looks about as easy as trying to write your own replacement for kernel32.dll 9_9

http://eternity.mancubus.net/pics/whoa.png

Just catching up here. I'm glad *something* is happening. Why not open it up to the world and let the hive mind do some of the clean up?

Jon said:

Why not open it up to the world and let the hive mind do some of the clean up?

I was asking the same question until I saw this. Whatever Quasar wants to do to prepare the database, it is probably happening very slowly. :Z

Xeriphas1994 said:

I was asking the same question until I saw this. Whatever Quasar wants to do to prepare the database, it is probably happening very slowly. :Z

Yeah. At this point it would be useful to have some help. I think we need to let some trusted editors into the "playpen", so to speak, and let them identify issues that need to be fixed, and I am already aware of a *number* of issues. I just don't have a lot of time at the moment, and Manc and I have trouble being in the same place at the same time with our schedules, so things go slowly.

It's very close to being ready so I guess we just need to buckle down and finish it off.

One thing I am very unsatisfied with is the fact that the Monaco skin doesn't display the licensing terms anywhere on it. As is, you must click through a sequence of 3 or 4 specific articles to find the license of the wiki, and IMO, that doesn't meet its requirements. The problem is, to get a logo or at least text onto every page to reflect the license requires modification of the skin itself, which is a VERY large and dense php script. I barely know anything about how the skin works. If anybody thinks they can pull this off, I'd be appreciative of their help.

I'd volunteer to help, but I don't like Monaco. :p