Jump to content

Wikipedia:Bots/Noticeboard/Archive 10

From Wikipedia, the free encyclopedia
Archive 5Archive 8Archive 9Archive 10Archive 11Archive 12Archive 15

Mass AWB edits that clutter up watchlists

Illustration of my watchlist.

Hi! Magioladitis (talk · contribs) has been mass-editing lots of WikiProject templates using AWB. I asked him to stop and he did.

I don't have much background about how bots or computer-assisted edits should be done on Wikipedia, hence I'm posting here. Is there a process that should be followed for making these kinds of automated edits? It's very disruptive for people like me who have lots of articles on their watchlist, see the screenshot.

If this is the wrong place to discuss this incident, please let me know. -- intgr [talk] 16:36, 3 November 2015 (UTC)

intgr I stopped. A bot, BattyBot already is doing something similar and I am still waiting for some stats to ask whether a bot can do it for me. My aim is to reduce the number of non-standard redirects of project banners. -- Magioladitis (talk) 16:39, 3 November 2015 (UTC)
I know that disruption of watchlists is a bad side-effect and I apolosise for this, but me and some other bot owners have to load a huge list of redirects everytime we want to tag talk pages. If we once go and clean those that do not start with WiKiProject... and/or WP... we can be more effective. -- Magioladitis (talk) 16:41, 3 November 2015 (UTC)
Every time this happens he promises to stop and then starts again shortly after. I'm starting to wonder whether a ban against all automated edits might be needed here. — Martin (MSGJ · talk) 16:42, 3 November 2015 (UTC)

MSGJ I did this User talk:Scott#WikiProject_template_redirects hoping for help on that and less disruption. As you may see we even lack the method avoid brute force in some cases. I even asked for a AWB FR: T117365. -- Magioladitis (talk) 16:47, 3 November 2015 (UTC)

@GoingBatty: -- Magioladitis (talk) 16:47, 3 November 2015 (UTC)

This weekend I was running BattyBot task 21 to add {{WikiProjectBannerShell}} to some talk pages, which has the side benefit of reducing the number of non-standard redirects of project banners. I also made some non-bot changes to help with this effort (with and without AWB), with the goal of only saving the edit if it would have a noticeable effect on the rendered page (e.g. adding a new parameter). GoingBatty (talk) 17:22, 3 November 2015 (UTC)

Before the statistics I was working blindly. After MSGJ's complains,10 I moved only to the non-standard names. I can stop the task completely. -- Magioladitis (talk) 16:50, 3 November 2015 (UTC)

@GoingBatty: Thanks. I am good with that. Please keep it up. -- Magioladitis (talk) 19:04, 3 November 2015 (UTC)

For information of interested parties, and related to the discussion above, I have today blocked Magioladitis for a week for repeated violations of the bot policy and AWB's rules of use. I welcome feedback on this action. — Martin (MSGJ · talk) 13:29, 17 November 2015 (UTC)

Turning off Hazard-Bot

Can Hazard-Bot be turned off, with regard to its continued cleaning out of the template sandboxes? It's been continuing its chore (certainly in X9, and I assume the others) for several days, twice a day, because Cyberbot-I is now cleaning, as noted previously. Dhtwiki (talk) 04:21, 26 October 2015 (UTC)

It looks like user talk archiving frequency may be too high given his current activity. I sent him an email. @Cyberpower678: The easiest fix for now is to get Cyberbot to use the same replacement text as Hazard-Bot. H-B doesn't have a per-task shutoff option and I don't want to block it given that it handles other things. — Earwig talk 04:32, 26 October 2015 (UTC)
It would be nice to retain Cyberbot-I's message, as well as schedule, as its message mentions the cleaning frequency, whereas Hazard-Bot's does not. I've asked to include the time as well, but one thing at a time. Dhtwiki (talk) 05:17, 26 October 2015 (UTC)
Sorry for lack of clarity. I was referring specifically to these; Cyberbot should be using {{subst:Template sandbox reset}}, but it's not. As for the template talks, I repeat my suggestion from earlier of creating a {{Template talk sandbox reset}} or similar. It may be easiest right now (although a little counter-intuitive) to create that template and then do a namespace-switch on {{Sandbox reset}} to substitute {{Template talk sandbox reset}} when it's in that namespace. Cyberbot should be tweaked to use that template, and Hazard-Bot should continue to work correctly assuming it works by substituting {{Sandbox reset}}. — Earwig talk 07:01, 26 October 2015 (UTC)
I don't see how getting Cyberbot to use a slightly tweaked message/substing templates is going to fix Hazard Bot's aggressive schedule. As I see it I only need to remove a space from Cyberbot to keep the messages consistent, and HazardBot needs to relax.—cyberpowerTrick or Treat:Online 19:17, 26 October 2015 (UTC)
It prevents the bots from fighting until we can get Hazard-SJ to tweak the bot. — Earwig talk 20:54, 26 October 2015 (UTC)
I removed the extraneous space from the message so the fights should stop anyways. As for Hazard-SJ he doesn't seem to respond to anything bot related if he's being asked to change it. What other things does HazardBot do?—cyberpowerTrick or Treat:Online 13:35, 27 October 2015 (UTC)
There's also the template talk pages though, right? As for the last point, it might be worth going to through Special:PrefixIndex/Wikipedia:Bots/Requests for approval/Hazard-Bot if we seriously intend to shut it off. These are happening actively but I'm not sure what else. — Earwig talk 03:15, 28 October 2015 (UTC)
(reply to The Earwig's first comment in this section) I just dug into GitHub (the GitHub link in the infobox on the bot's user page seems to be an old repository, but I found what looks to be the current one) and Hazard-Bot's userspace, and it appears that there is an individual shutoff for the sandbox cleaning task. See User:Hazard-Bot/DoTask, User:Hazard-Bot/DoTask/SandBot, and the code here (in particular, line 40 and lines 132 to 151). jcgoble3 (talk) 05:57, 28 October 2015 (UTC)
Thanks for that. Disabled. — Earwig talk 06:06, 28 October 2015 (UTC)
Sorry for the delay, I'm horribly busy IRL, but the frequency of the task was actually every hour (the main sandbox included; I just changed it to once a day), although it has checks that will delay and eventually skip if it recognizes that the sandbox is in use. As for the content, yes, as the code linked above shows, I do substitute via the template, but another thing to point out is that the code actually checks to make sure it isn't editing if a sandbot was the last to edit, except that since Cyberbot I was only recently approved for the task, it was not being recognized as a sandbot (now it is). As far as I'm aware, the task should be fine to re-enable, if you guys are okay with it (once again, the frequency was changed, and if Cyberbot I for some reason picks a new message of it's own, should either way get skipped, like all the other [recognized (in my code)] sandbots.  Hazard SJ  06:09, 19 November 2015 (UTC)
Looks fine to me. Re-enabled. We can deal with a new sandbox template at some point in the future if people are set on changing it. — Earwig talk 06:21, 19 November 2015 (UTC)

Lowercase sigmabot II down; sandbox header not being immediately reinserted

It seems Lowercase sigmabot II is no longer operating; the bot is best known for its ability to immediately reinsert the sandbox header in the event that it gets removed; no other bot that cleans the sandbox, to my knowledge, does this. Can this functionality be incorporated into Cyberbot I? Just wondering. —k6ka 🍁 (Talk · Contributions) 18:00, 25 November 2015 (UTC)

I just tested this, and it looks like it's working fine now. — kikichugirl oh hello! 05:05, 26 November 2015 (UTC)

Too many sections on User talk:Andrzejbanas

Can you move sections (February 2014 - present) into archives? Because it's too many. 183.171.178.188 (talk) 14:34, 17 January 2016 (UTC)

 Not done Nothing to do here, there are no bots set up to archive that page. — xaosflux Talk 15:11, 17 January 2016 (UTC)

Requesting apihighlimits flag for AnkitAWB (talk · contribs)

Well, it's bit hard to go through categories one by one, it'd be better if I can load a category at once and roll on it. For what it's worth, my bot's had a mentionless history and has registered 2 complaints throughout it's running time and they were not even his fault (caused by putting the articles in the wrong stub category). Also, I've made a few changes, now it integrates with BannerShell and fixes alternate templates (Custom Module) instead of skipping them. --QEDK (T 📖 C) 09:52, 18 January 2016 (UTC)

Also, when I filed my BRFA, I put in a value of 3 epm. Now, I've tried to clock my bot between 6 - 12 epm generally (it's working right now at 7 epm). I hope that's not too much of an issue. Also, what's the general roof limit? --QEDK (T 📖 C) 11:32, 18 January 2016 (UTC)
@Xaosflux: I think I got an error dialog box when I tried to load pages using the plugin (NoLimitsPlugin.dll). I'll try again on its next run. I don't think that epm will be an issue because well, I'm running the bot with a 0 second delay and it hasn't capped 14 epm. Thanks a lot. --QEDK (T 📖 C) 13:33, 18 January 2016 (UTC)
Seems to be working fine. Probably logged in from wrong account. --QEDK (T 📖 C) 06:41, 21 January 2016 (UTC)

Archivebot?

This is getting quite embarassing too:

... failing (and still awaiting) an adequate response by the botop in question: are there any other bot owners that can explain a bit (at Jimbo's talk page preferably)? --Francis Schonken (talk) 12:06, 26 January 2016 (UTC)

SineBot kaput

SineBot (talk · contribs) has not signed anything since last Friday. I have left a note on Slakr's talk but don't know if he's around to reboot it - can anyone else get hold of him? I'm sure I've said this before, but since this is such an essential tool, shouldn't the WMF be taking care of it? Ritchie333 (talk) (cont) 10:15, 26 January 2016 (UTC)

It's been down a lot recently, not sure why. You can also try pinging him on IRC. —k6ka 🍁 (Talk · Contributions) 13:42, 26 January 2016 (UTC)

Magioladitis

Magioladitis is blocked again.I have opened a discussion at Wikipedia:Administrators' noticeboard#Magioladitis to explore ways to resolve the issues. Anyone is invited to comment there. — Martin (MSGJ · talk) 20:47, 28 January 2016 (UTC)

Murph9000 Bot

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.


Is Murph9000 Bot (talk · contribs) an approved bot? There is no entry "changed group membership for Murph9000 Bot from (none) to bot" at their logs, and I can't find a WP:BRFA. --Redrose64 (talk) 09:45, 1 February 2016 (UTC)

  • It is not an approved bot. According to the tag on its userpage it's used for WP:ASSISTED editing, not fully automatic bot editing. The template removal appears to be somewhat quick, though; pinging @Murph9000: about this.Jo-Jo Eumerus (talk, contributions) 09:53, 1 February 2016 (UTC)
    • It was an WP:ASSISTED edit. Every single edit was reviewed and approved manually, with regular spot checks of a significant number of the diffs post-edit (both while it was ongoing, and at the end of it) to ensure that they were as desired. I'm happy to answer any questions. Murph9000 (talk) 09:57, 1 February 2016 (UTC)
      • If "every single edit was reviewed and approved manually", why was there any need for "regular spot checks"? Wouldn't you have already reviewed every one? Anomie 01:54, 2 February 2016 (UTC)
        • Precisely the same reason why I routinely review diffs after editing without any assistive tools. To confirm that the change I intended to save actually applied correctly. It is one thing to review a pre-save diff (or preview), and a quite different thing to review a post-save diff. Regardless of the interface or tools used, I do not solely rely on checking prior to saving, because MediaWiki has a non zero failure rate at the point of saving. In the case of removing those templates, it was trivial to tab focus between windows, refresh the contribs page, mass-check the size of 100% of the diffs, and spot check a few of the diffs in detail. Being a small change, I could review several post-edit diffs to confirm intended changes in just a few seconds. Using the standard web interface, without any extra tools involved, I've had multiple instances where MediaWiki's server-side edit conflict detection has failed, so I always try to pay attention to actual changes after the save is complete. Occasionally, the actual change does not match the change which was intended. Technology is fallible. Why would I take any less care, or pay any less attention to post-edit checking when using assistive tools? Per WP:MEATBOT (the preamble to WP:ASSISTED), I was taking maximum care to pay attention to the edits I was making, ensuring that the net result was no less reliable than an unassisted normal edit. The pre-save review could reliably be performed quite quickly for the diff of a 27 byte change involving a single line (with decades of experience of reviewing Unix diffs), prior to approving the change. That left plenty of time (while waiting on the save to complete, and the next change to be shown and prompted for approval) to use another window in parallel and double check things. I am responsible for the edits that I make, and I take that very seriously. The habit of checking and then double checking diffs pre-dates the existence of MediaWiki by multiple decades. It pre-dates the web itself! I did it quickly, but with considerable care and attention. After I save this edit to add this message, my next and immediate action will be to go to the page history and check the diff for it. Murph9000 (talk) 03:07, 2 February 2016 (UTC)
As a bot operator myself I see two troubling concerns:
  1. Usernames which could be easily misunderstood to refer to a "bot" (which is used to identify bot accounts) or a "script" (which alludes to automated editing processes), unless the account is of that type. from WP:IU
  2. A bot process operating without getting a BRFA and documentation as to what tasks it's doing (and where the consensus/authority for such a change stems from).
  3. Even assisted edits still require a BRFA due to the nature of fait acompli issues.
For these reasons I suggest that the bot be blocked until Murph9000 completes a BRFA (along with documenting the consensus for the task) and properly Bot-izes the task. Hasteur (talk) 13:24, 2 February 2016 (UTC)
From OP's explanation above, my only issue would be having "Bot" in the name and not actually being a bot. As long as every edit is checked, then it's pretty much like using any other assisted tool/script. I think blocking and demanding BRFA for this many pages is way too harsh, considering we have editors making edits in batches of hundreds without any BRFAs or secondary accounts. To be safe, we can speedy approve an assisted editing account if the OP wishes to pre-state their intentions. —  HELLKNOWZ  ▎TALK 13:33, 2 February 2016 (UTC)
@Hasteur and Hellknowz: If I misinterpreted the username policy, I apologise for that. My interpretation was that it was actually required to be "… bot", due to WP:ASSISTED (aka WP:BOTASSIST) falling within the bot policy. I'm happy to request rename, but will not do so without clarity on that issue. I feel that the policy, as written, requires that naming for an alternative account used or assisted editing. I'm not entirely sure what you would like from me in terms of an "assisted editing account" request. Do you mean to submit a bot approval request specifying WP:ASSISTED, or something else? My intent was (and is) to occasionally use semi-automated (with manual review and approval of individual changes) tools to perform occasional ad-hoc WP:ASSISTED editing, on a small scale; while avoiding cases which do not appear to be supported by consensus or appear to have possible controversy. Never fully automated. Never unattended. It felt borderline whether I actually needed to use an alternative account, but I chose to do so for transparency, taking care to properly link it an place the suggested notices, and wanting to avoid any chance of being accused of trying to operate an unapproved bot. The nature and scale of the intended task felt safe and small enough to not need formal approval even as an assisted edit. Murph9000 (talk) 15:34, 2 February 2016 (UTC)
Enacting the outcome of a TFD is a normal bot task (and I think there is a bot already authorized to do that). Post facto stylizing of your user signature via a bot is against WP:COSMETICBOT policy, and outside of bot involvement should only be done in the cause of some other change. For these reasons, I reiterate that this user account should be blocked from making any edits until a BRFA has passed and there's a consensus for the bot to do work. Hasteur (talk) 16:30, 2 February 2016 (UTC)
(edit conflict)I mostly agree. If it's not an approved bot, the username policy states that it should not have 'bot' in the name. At the same time, I know I've never "spot checked" my contribs for accuracy unless it was a bot run. It's running consistently about 1 edit per 10 seconds. On the 22nd, we see the bot account make 158 edits consistently at the pace of about one every 20 seconds for 38 minutes without a slowdown or break. It marks its edits with "Bot:" [1]. Blocks are not punitive, and the unapproved process does not seem to have been running for almost a week now, so I would suggest there is no real need to hastily block at this time. I would suggest that @Murph9000: seek approval / trial via BRFA if they would like to make automated edits. If you are just clicking 'save' blindly, and "spot checking" them later - I would argue that this is really no different from having a script do it for you. SQLQuery me! 13:43, 2 February 2016 (UTC)
@SQL: As I explained, I routinely check my edits post-save, no matter what tools are involved, or if it's a simple and entirely unassisted browser edit. It's a decades old habit, and it will likely never change, as it has served me well over the years. I consider not doing such checks to be a bit reckless. The 10 second pace is trivial to explain. I had it throttled to that rate to avoid creating excessive server and database load spikes, as well as to avoid me slipping into a blind rapid fire mode where something unintended would be more likely to slip through. The precise nature of the edits I was making meant that I could safely and reliably review and approve each pre-save diff in around 1–2 seconds (I was performing those diffs client-side, so no network round trip to add delay). For a more complex change, I would certainly take longer to approve each change. I actually spent a fair amount of time patiently waiting on the throttle timer (I only tabbed to the contribs window to spot check once every few edits, after queuing up a couple of saves. If I had run at roughly maximum rate (still with full manual approval, but without checking post-save diffs in parallel), I could have easily and safely completed the task in around 15 minutes (if the server accepted the changes at that rate). I most certainly was not "blindly" approving the changes, there was just very little time required to actually look at them with a highly experienced eye. If you look closely, you should see some gaps in the timing where I paused to take a drink or respond to something else, as I distinctly remember there being a number of points where I paused and the throttle timer had expired (some cases longer than others). I am more than capable of working solidly at a task without a break for longer than 38 minutes, when I have my tools properly setup and set myself about getting the job done.
I was simply trying to help out another editor (who asked for assistance on WP:HD) with an ad-hoc task which seemed uncontroversial, small scale, looked to have consensus clearly established through proper process and without objection for around 48 hours post-closure. I am now starting to sincerely regret offering assistance, and trying to contribute to clearing one of the many backlogs.
I have no interest in operating a fully automated process at this time. There is no specific process to approve, no trial to perform, as it was an ad-hoc assisted edit session. It was never my intention to run a fully automated process. If that changed, I would certainly seek prior approval.
Murph9000 (talk) 15:34, 2 February 2016 (UTC)
So, then per the username policy, that account should be blocked, as it is unlikely to ever be a bot account? SQLQuery me! 15:46, 2 February 2016 (UTC)
I think the OP means (per reply to my comment) that they thought it had to be "Bot" even for assisted editing, because assisted editing falls under BOTPOL too. —  HELLKNOWZ  ▎TALK 16:00, 2 February 2016 (UTC)
Throwing all the other odd coincidences out the window - the OP speaks as well of somehow 'throttling' their 'manual edits'. How exactly is this accomplished? Why does the 'editor' in question mark edits as 'Bot:'? What platform are we using to accomplish these 'manual edits'? SQLQuery me! 16:04, 2 February 2016 (UTC)
There's no way an account could ever mark edits as bot edits if they don't carry the bot flag.—cyberpowerChat:Limited Access 16:43, 2 February 2016 (UTC)
Unless they are in trial ;) —  HELLKNOWZ  ▎TALK 17:12, 2 February 2016 (UTC)
Ok, I find myself in the situation where I feel quite significantly offended and insulted by a combination of certain remarks made here; particularly those remarks which which basically indirectly accuse me of lying, and in a generally hostile and bad faith tone. Much of the conversation was kept perfectly reasonable, and I thank you for those portions of it. I do not see anything useful being achieved by trying to convince particular individuals who have clearly jumped to their own false conclusions and whose language leads me to believe that they are extremely unlikely to change their minds. I simply do not want any more stress from this, and it has already consumed an absurd amount of text over a successful good faith change which totalled 4,266 bytes. I apologise for anything I did that was wrong. My intentions were always to try to be fully transparent and stick to the published policies, which I believed I had done. I only ever intended to try to make a useful good faith contribution that was consistent with policies, guidelines, and consensus. As far as I am aware, the WP:ASSISTED edits I made were 100% successful, precise, and did what needed to be done without causing undue disruption. I will direct my efforts towards something else which does not cause me stress and offence, most likely outside the Wikimedia Foundation (that is not to say that I am retiring from WP, only that I'm scaling back my efforts, for a while at least). I have no desire to provide voluntary effort and assistance where it is not wanted. Murph9000 (talk) 17:27, 2 February 2016 (UTC)

Looks like I came here too late to change anyone's mind, but I was just going to say that you could easily rename the account to something like "Murph9000 (Assisted)" and be done with it. The edits do not look problematic. The sig-cleaning wouldn't be an appropriate bot task, of course, but as only 30 edits done using some (admittedly unspecified, but I'm going to guess pywikibot) assisted program is not worth fretting over. Seriously.

  • @SQL: I only see two edits marked as "Bot:", which looks to be an archive script; there's nothing against running that in your userspace. For the record, I spot-check my edits all the time—my bot's edits too—I don't see why anyone would contest that.
  • @Hellknowz: Cyberpower was referring to the bot RC flag, which wasn't involved here; i.e. the technical feature as opposed to putting the phrase "Bot:" in an edit summary. Of course, anyone can do the latter, but shouldn't outside of their userspace unless trialed/approved—which wasn't violated here.

— Earwig talk 19:50, 2 February 2016 (UTC)

@Hellknowz and SQL: Who is "the OP"? Nobody contributing here seems to have those initials. --Redrose64 (talk) 20:04, 2 February 2016 (UTC)
Sorry for confusion, I meant OP as in "operator", a shorthand I've gotten used to in BRFAs. (I realize it conflicts with "original poster".) —  HELLKNOWZ  ▎TALK 20:14, 2 February 2016 (UTC)
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Breaking change

In a month or two, they're changing the login system for bots. You can read more here: phab:T121113 and http://lists.wikimedia.org/pipermail/wikitech-l/2016-January/084501.html

If you expect any problems, please reach out to mw:User:Anomie or mw:User:Tgr (only you have to be nice to Tgr, because he did me a favor last week.  ;-) This was in Tech/News today, and it's easy to miss an announcement like this, so please also check in with your friends and other wikis and make sure that none of them will be surprised by this. Thanks, Whatamidoing (WMF) (talk) 22:01, 18 January 2016 (UTC)

Looks like lazy botops (me) will just use Special:BotPasswords and fancy botops can transfer to OAuth (already possible to start using after enabling it, as far as I understand). Tool makers with logins (hi, AWB) will have some work to do as they will need at least some OAuth support. —  HELLKNOWZ  ▎TALK 22:25, 18 January 2016 (UTC)
OAuth is easy enough to set up, especially considering bot ops are meant to have some experience with technology... ;) even I managed to do it the other day! Mdann52 (talk) 22:31, 18 January 2016 (UTC)
Having experience with technology, and unlimited free time to deal with these changes are two very different things. SQLQuery me! 15:52, 22 January 2016 (UTC)
I don't have much time myself, but I managed to setup OAuth for my bots. It also gave me the opportunity to further secure my bot accounts, to prevent unauthorized access. Peachy uses OAuth by default now, and anyone using it will likely have discovered that I have forced the usage of OAuth on that framework, so for them it's a matter of giving Peachy the 4 tokens it needs to run.—cyberpowerChat:Online 15:56, 22 January 2016 (UTC)
That's one way to look at it, I suppose. Anything where I'm going to have to dive in and mess with 10+ year old legacy code isn't really as simple as it sounds, however. SQLQuery me! 16:00, 22 January 2016 (UTC)
OAuth was a bit difficult to setup in Peachy. That code is old. But the underlying mechanics is pretty simple to understand and implement.—cyberpowerChat:Limited Access 16:05, 22 January 2016 (UTC)
Thanks for the heads-up! — xaosflux Talk 22:48, 18 January 2016 (UTC)

My bot broke yesterday ("There was an unexpected error logging in."). I presume this is related? —Steve Summit (talk) 13:06, 22 January 2016 (UTC)

Same. Apparently, I get a response for needing token after sending the first token. I set up a Special:BotPasswords and it's the same. So I tried to pass the new token and it just asks for another (probably repeating this forever). It sounds like something like a bad cookie container would cause, but I am using a proper one (as far as I know), so I don't know. I might just set up OAuth to avoid all the hack-around. —  HELLKNOWZ  ▎TALK 13:34, 22 January 2016 (UTC)
This is a common issue when switching login methods. Delete the cookie the bot uses. That should fix you problem in most cases.—cyberpowerChat:Online 13:58, 22 January 2016 (UTC)
PS switching to OAuth will likely have the same issues in regards to token retrieval. You need to delete the existing cookie the bot uses. Then your bot will work if you switched to a new method.—cyberpowerChat:Online 14:06, 22 January 2016 (UTC)
That's the thing -- there are no cookies stored between running the bot, I create a new container each time, for each run. It sure looks like a cookie issue, but I can't figure out what. —  HELLKNOWZ  ▎TALK 14:14, 22 January 2016 (UTC)
Me, too. —Steve Summit (talk) 14:41, 22 January 2016 (UTC)
Feel free to ignore this if it is totally irrelevant, but seeing that these are very recent login problems, could they be related to Wikipedia:Administrators' noticeboard/Incidents#No longer able to log out (thank you, WMF!)? Fram (talk) 14:27, 22 January 2016 (UTC)
Funny you should post that, as that happened to me while switching accounts and I was like "oh, come on" before just logging in the bot in a different browser. —  HELLKNOWZ  ▎TALK 14:37, 22 January 2016 (UTC)
It does appear that a cookie "forceHTTPS=true" is being given. I'm not using a secure connection, so that may be why it fails. I wish the API actually said something to that effect instead of endless "NeedToken" responses. —  HELLKNOWZ  ▎TALK 14:37, 22 January 2016 (UTC)
Yep, that seems to have done it. Switching to https passed the login as normal (with the new bot passwords thing). I can't find any place where I am specifying that I want to force the account to always use https, so it's probably some logic in recent updates that disabled non-https. —  HELLKNOWZ  ▎TALK 14:56, 22 January 2016 (UTC)
I thought Wikipedia moved to HTTPS ages ago and using simple HTTP was highly discouraged.—cyberpowerChat:Online 15:47, 22 January 2016 (UTC)
Thanks for posting about that, Hellknowz. It led us to spot that the cookies were being served with the "secure" flag even if the API was being accessed over plain http, which was apparently causing problems for a lot of people. At some point we should fix it so the API requires https like index.php does; I had thought it was that way already. BJorsch (WMF) (talk) 13:51, 28 January 2016 (UTC)
  • On the topic of being logged it, I find my self persistently getting logged out. It's rather annoying. I almost typed a response here logged out.—cyberpowerChat:Online 15:47, 22 January 2016 (UTC)
  • Oh dear Unicorn. This means I have to figure out how to login with my bots again. It's been years since I've done that. Headbomb {talk / contribs / physics / books} 18:06, 22 January 2016 (UTC)

How is it even possible to login any more without HTTPS? See m:Talk:HTTPS#Bots for the scrambling I had to do after my bots were knocked offline by the mandatory switchover to HTTPS. Wbm1058 (talk) 18:15, 27 January 2016 (UTC)

This highly technical discussion is mostly over my head, and I have no idea whether I need to change anything to keep my bots running. So far they seem fine and unaffected. This is not a good way to notify me, as I don't even have this page on my watchlist. I just happened to stroll by because of #Archivebot? below. If someone would take a glance at my bots' code and tell me if I need to change anything and if so, what I should change, I would be grateful. Thanks, Wbm1058 (talk) 15:49, 27 January 2016 (UTC)

It's relatively simple. The explanation of the change is complex. You have 2 options before your bots will break. The first option is to implement OAuth, and have your bots interact with the API with OAuth. This method is more secure and should avoid any future breaking changes, including this one which will be deployed once the security bugs have been fixed. Your second option is to wait until your bot breaks, which will also be the time when Special:BotPasswords becomes available for use, where you create a password for your bot, and then change your bot's login credentials accordingly.—cyberpowerChat:Limited Access 17:00, 27 January 2016 (UTC)
Off the top of my head, I don't know what "OAuth" is, much less how to implement it. My bot already has a password (another name for "login credentials"). Surely "Special:BotPasswords" wouldn't be a page where administrators can go to look up all the bots' passwords? "The explanation" still needs to "use English" ;) Wbm1058 (talk) 17:20, 27 January 2016 (UTC)
No, BotPasswords is a place where you can create new type of password for your bot, and can also place restrictions on it (such as 'read only' or only from certain ip address). You can look at the interface at testwiki; you can NOT retrieve other peoples passwords with it. — xaosflux Talk 17:32, 27 January 2016 (UTC)
Thanks, that's helpful. So once I use this interface to set up a new PW, it should just work with no changes to my bot's PHP code other than changing the values of password constants to the new password? Are there any special considerations for running the bot from multiple computers, say, for redundancy? What if I want to run the bot from a laptop with a mobile connection? Wbm1058 (talk) 18:24, 27 January 2016 (UTC)
I don't know the answers to most of those questions, but one subtlety you should be aware of is that, if you use "Special:BotPasswords", the "username" value you send to the login action will be of the form "somename@somethingelse" and will not be the same as your bot's "username" as displayed in Wikipedia. --R'n'B (call me Russ) 22:02, 27 January 2016 (UTC)
@C678: Normal logins shouldn't break yet (if anything breaks, it's either bugs in bots' cookie handling or bugs in the new SessionManager code that need fixing). The rollout of OAuth owner-only accounts and of BotPasswords is intentionally being done sooner to give people more time to change over before AuthManager does start breaking things. BJorsch (WMF) (talk) 13:51, 28 January 2016 (UTC)
@Wbm1058: FYI, it was also announced on mediawiki-api-announce, wikitech-l, and in the Tech News posted to WP:VPT. BJorsch (WMF) (talk) 13:51, 28 January 2016 (UTC)
Thanks. I don't subscribe to any of the email lists; my email is already a spam-disaster and I don't open 90%+ of the emails I receive (and I'm talking about those that do make it through Yahoo's spam filter). I do look at the Tech News, but not religiously, and that doesn't do a great job of differentiating the important changes from the trivial in a way that catches my attention. So, glancing at the emails I see more jargon that I don't understand. Can you disambiguate the term Consumer (disambiguation) for me? I don't know what that means in this context. Wbm1058 (talk) 14:33, 28 January 2016 (UTC)
In OAuth, the program that uses OAuth to connect to a website is called a "consumer". Anomie 13:17, 29 January 2016 (UTC)
  • I find myself fighting very hard with OAuth...now that everything is installed, I keep getting:
Error

tools.deltaquad-bots@tools-bastion-01:~/pywikipedia/core$ python pwb.py login -oauth OAuth consumer key on en:wikipedia: 4d062016815b03ef6a09dea8e62443a5 OAuth consumer secret for consumer 4d062016815b03ef6a09dea8e62443a5: Logging in to wikipedia:en via OAuth consumer 4d062016815b03ef6a09dea8e62443a5 ERROR: Expected x-www-form-urlencoded response from MediaWiki, but got something else: 'Error: "DeltaQuadBot Labs Authentication" is an owner-only Connected App. To fetch the access token, see Special:OAuthConsumerRegistration/update/4d062016815b03ef6a09dea8e62443a5.\n\n<span class="plainlinks mw-mwoautherror-details">Consumer is owner-only, E010</span>' ERROR: Access token not set ERROR: Invalid OAuth info for wikipedia:en.

Is usage on tool labs require the consumer to be non-owner only? -- Amanda (aka DQ) 17:37, 4 February 2016 (UTC)
In an owner only consumer, you should already have the access token and secret. You need to feed the bot with those keys too.—cyberpowerChat:Limited Access 17:49, 4 February 2016 (UTC)


Bots broken again

I moved my bots to use OAuth, to avoid any breakage in the change. Well now they're broken, and I don't think it's my fault. Using OAuth I get persistent badtoken error regardless of how many times I delete the bot's cookie. I have made no modifications, and other bot scripts that were working are now broken too.—cyberpowerChat:Offline 04:57, 5 February 2016 (UTC)

We talked about this on IRC yesterday, and it turned out to be a bug in the (non-SessionManager) handling of sessions in the OAuth extension when Setup.php had already started a session due to a session cookie. For Cyberpower678, it was worked around by restarting his bot to clear the bad cookie. The fix for the extension is gerrit:268702, although SessionManager coming back in wmf.13 should fix it too. Anomie 12:37, 6 February 2016 (UTC)
I knew I should've made the "have you tried turning it off and on again?" joke... —  HELLKNOWZ  ▎TALK 12:53, 6 February 2016 (UTC)

Bot user rights at Metawiki

  • According to mailarchive:wikitech-l/2016-January/084501.html, I should log in as my bot and go to m:Special:OAuthConsumerRegistration/propose in order to switch my bot from password to OAuth mode. However, that page doesn't work since my bot account isn't autoconfirmed on Metawiki. This looks like a stupid restriction since most bots probably only make edits to one project but still need to visit that page. Will the bot account automatically become autoconfirmed in a few days, or is mw:Manual:$wgAutoConfirmCount set to a nonzero value on Metawiki? --Stefan2 (talk) 22:02, 9 February 2016 (UTC)
    • When I first went to the page with the bot account, it said I needed auto-confirmed, because I had never visited meta:. I just logged in to check again and now it shows the full page. So it seems just waiting the day limit is sufficient, though annoying. —  HELLKNOWZ  ▎TALK 22:10, 9 February 2016 (UTC)
      • A brilliant idea would be to go to the OAuth form, and directly login to the bot there to create the owner-only consumer, without actually logging out from your own account, that way you can easily setup OAuth without being blocked from the restrictions.—cyberpowerChat:Online 23:14, 9 February 2016 (UTC)

And again...

Question regarding notices to User talk pages

A request was posed on my user talk page to not have the bot clobber pages that are redirects when delivering the friendly notices for Wikipedia:Bots/Requests for approval/HasteurBot and Wikipedia:Bots/Requests for approval/HasteurBot 2. In the past I made changes that were reasonable to accomidate similar changes (ex: following the redirect and delivering the notice there) only to have some editors claim that I have exceeded what the bot was authorized to do so I'd like to ask the question here: Do I need to file a new BRFA to change the behavior in edge cases dealing with pages that hard redirect to a new location? Thank you for your consideration. Hasteur (talk) 18:03, 21 February 2016 (UTC)

I usually wouldn't consider minor changes involving edge cases to require a new approval (clarification here or on a similar forum should be fine). In this case, following the redirect seems like the natural option, and it's how I've handled it when I've implemented this behavior; why were people opposed to that? — Earwig talk 19:31, 21 February 2016 (UTC)
They weren't necessilary opposed to the change I did, but (in my viewpoint) they had a finely ground axe that they wanted to embed in my wiki-career. Hasteur (talk) 20:50, 21 February 2016 (UTC)
Removing the redirect is not appropriate; either skip or follow the redirect, agree these are edge cases and do not need a new approval. — xaosflux Talk 03:18, 22 February 2016 (UTC)

Legobot misbehaviour

Legobot (talk · contribs) is misbehaving, and is listing discussions that are either closed RfCs, or not RfCs at all. I have posted to its talk page, but with no response from Legoktm (talk · contribs); and since the bot continues with incorrect edits, should I block the bot? --Redrose64 (talk) 00:10, 7 March 2016 (UTC)

Redrose64 Are the only bad edits on Wikipedia:Requests for comment/Unsorted and Wikipedia:Requests for comment or is it also making bad edit elsewhere? — xaosflux Talk 00:16, 7 March 2016 (UTC)
If so, I suggest slapping a {{bots|deny=Legobot}} on those pages and giving the operator a bit more time to reply. — xaosflux Talk 00:18, 7 March 2016 (UTC)
I emailed Legoktm as to this discussion as well. — xaosflux Talk 00:20, 7 March 2016 (UTC)
Uhh, this is weird. I haven't changed the code at all in a few months, plus I don't have time right now to look into it (maybe tomorrow?), so I can turn off that part of the bot if it's causing widespread problems...is it? Legoktm (talk) 07:12, 7 March 2016 (UTC)
I got Legobot to behave properly with the non-RfC at WP:VPR by using redirects that don't begin {{Rfc}}, so that Legobot doesn't think that it's an open Rfc. --Redrose64 (talk) 20:15, 8 March 2016 (UTC)

You may need to restart your bots twice

Hi everyone,

Ops is planning a major data center migration during the week of 21 March. From the POV of a bot owner, there are two major effects:

  • Your bot will not be able to edit for approximately 15 to 30 minutes at the start and end of the main action, on Tuesday, 22 March or Thursday, 24 March. [1]
  • Some bots may need to be re-started after the read-only window ends.

This is being announced in m:Tech/News and through the wikitech-l mailing list. phab:T124671 seems to be the central task. You can follow those for further updates (or {{ping}} me). Please share this information with other projects and bot owners.

  1. ^ The exact time has not been chosen yet, but my guess is approximately 07:00 UTC. Also, as with all tricky operations, it may be postponed for safety's sake; in that case, the next probable time window is the third week of April.

Whatamidoing (WMF) (talk) 16:28, 10 March 2016 (UTC)

I'm a little confused as to why the bot needs to be restarted. Won't the API simply return an error during the period, or do we have to deal with bad session data again?—cyberpowerChat:Limited Access 17:14, 10 March 2016 (UTC)
Yes, the API will return an error. But some bots are a little brittle about this, and the error may not be the same one that we "usually" see (read-only mode isn't the same as the 'blue screen of death'). Whatamidoing (WMF) (talk) 18:15, 10 March 2016 (UTC)

Postponed until mid-April

Ops has just postponed this until the week of 18 April. Some backend services will migrate to codfw during the week of 04 April (maybe on the Thursday?). If your bot might rely on RESTbase, Parsoid, or other things like that, then {{ping}} me and I'll see if I can find a list.

I hope that everyone survived the five-minute test yesterday with no problems. Whatamidoing (WMF) (talk) 18:41, 16 March 2016 (UTC)

Hazard Bot

Wikipedia:Bots/Requests for approval/Hazard-Bot 33

How did this get approval without site-wide consensus such as from Village Pump? All the best: Rich Farmbrough, 20:07, 21 March 2016 (UTC).

See Wikipedia:Research help/Proposal#Project steps. VPT is in step 7. – Jonesey95 (talk) 20:13, 21 March 2016 (UTC)
Cart. Horse. All the best: Rich Farmbrough, 11:17, 22 March 2016 (UTC).
  • This approval really was quite premature. The template being placed on pages was outright nominated for deletion recently and it reached a "no consensus" outcome with significant support for deletion. There's not even a clear consensus that this template should exist on the project at all, let alone be added to thousands of articles. See the TfD itself. I have a strong feeling this process is going to leave everyone annoyed. Editors have invested large amounts of time into this proposal, botwork, reversion of the botwork, etc. before bothering to check if the community at large is ok with this. Why isn't this link just placed on the left sidebar for all articles? ~ RobTalk 02:18, 31 March 2016 (UTC)
  • I agreed to a small (100 edit) trial - just to have something tangible but easily removable to demonstrate to the community what was being attempted., The_Earwig was there other pages besides the ones link that led to ramping this up to 10000 pages? — xaosflux Talk 02:55, 31 March 2016 (UTC)
    • @Xaosflux: Just to be clear, I wasn't blaming you, Earwig, or anyone else handling approvals. I'm sorry if my message gave that impression. I'm hoping this can just be remembered as a sanity check for the future. Consensus developed in small groups really isn't sufficient when we're talking about basically adding something to MOS:LAYOUT. ~ RobTalk 03:09, 31 March 2016 (UTC)
  • Well, I mostly hate it, but that's beside the point. The idea here was a relatively small and short trial to get some attention and data, because we had basically twenty users expressing support, no opposition, and one neutral in the pre-bot discussions. The 100-edit trial passed with no real catastrophe and was too small to get wider attention. It's not dangerous and it can easily be blanked if notable opposition came about. My hope was to generate discussion, which has happened. — Earwig talk 03:40, 31 March 2016 (UTC)
    There was and is opposition, going back to 18 December last year [[2]]. And advice from User:Green Cardamom that they believe it would be healthy to voice other points of view outside those immediate parties who this is benefiting. And advice to take it to VP, which was ignored.
But the "easily blanked" was reverted. Oh well.
I raised the matter at VP(Proposals) and there is consensus to blank the template, until a VP proposal passes to utilise it more.
All the best: Rich Farmbrough, 00:39, 1 April 2016 (UTC).

User:AlexNewArtBot/ArchitectureSearchResult

This bot is causing spam to fill up "what links here". For example a new page created by me Columbjohn which has very little to do with architecture now has dozens of links to user pages due to this bot. Have never seen anything like it since 2010. Seems a totally useless bot to me, can it be stopped? I note the user has already been blocked. Thanks.(Lobsterthermidor (talk) 21:35, 1 April 2016 (UTC))

@Lobsterthermidor: That page is updated by User:InceptionBot which I operate. The "what links here" stuff will disappear about 14 days after an article is created. --Bamyers99 (talk) 22:51, 1 April 2016 (UTC)
Good! Thanks.(Lobsterthermidor (talk) 08:43, 2 April 2016 (UTC))

Rotten Tomatoes subtemplates

Theo's Little Bot (run by Theopolisme) has been recreating some of the subtemplates that were slated for deletion in this TfD. It's just been recreating them as an error. The bot owner hasn't edited for a few months, but is anyone able to disable only this task? It seems unreasonable to monitor the bot for when it decides to recreate one of these. Example: Template:Rotten_Tomatoes_score/0118688. ~ RobTalk 15:07, 19 March 2016 (UTC)

Template:Rotten Tomatoes score/1300854 as well. The bot only seems to be recreating the error pages, namely pages that will never be used anyways. -- Ricky81682 (talk) 22:37, 19 March 2016 (UTC)
I don't believe it's possible to disable one of a bot's tasks without blocking the bot entirely. I ran into this issue several months ago; the solution is to template-editor-salt them – see my rather brief protection log. – wbm1058 (talk) 22:49, 19 March 2016 (UTC)
Another editor contacted Theo offwiki to see about shutting down just task 22. In the meantime, I guess I'll just re-nominate as WP:G6 with a note to WP:SALT as they come about. Thanks. ~ RobTalk 22:58, 19 March 2016 (UTC)
  • FYI, the bot is also having issues with another one of it's tasks, resizing images. See User talk:Theo's Little Bot#Task 1. Elisfkc (talk) 03:59, 31 March 2016 (UTC)
    • That's a different problem. The Rotten Tomatoes problem is that the bot makes edits it's no longer supposed to do, which is problematic and lead to a block of the bot. The problem you mention is that the bot sometimes doesn't make edits it's supposed to do, which is less problematic (a backlog is generated, but it isn't as disruptive for the project and isn't a reason for a block). It would be nice if the bot operator could fix both problems, but only the Rotten Tomatoes problem needs to be fixed in order to unblock the bot. --Stefan2 (talk) 10:09, 31 March 2016 (UTC)
      • True, but it's something else that needs to be fixed with that bot. If there is an old version of the bot's code, maybe that will fix both issues. Elisfkc (talk) 15:52, 31 March 2016 (UTC)
        • Problem 1 (Rotten Tomatoes) is fixed if the bot operator removes a line from a cronjob (or whatever job scheduler the operator uses for this purpose). Problem 2 (file size reductions) is fixed if the bot operator corrects a bug in one of the bot's scripts. However, the problems can only be fixed by the bot operator, and the problem is that he isn't responding. --Stefan2 (talk) 21:29, 31 March 2016 (UTC)

@Xaosflux: I emailed Theo and he said the offending task has been disabled and the bot can be unblocked. I'll leave that decision up to your discretion. --Closedmouth (talk) 09:30, 4 April 2016 (UTC)

Closedmouth Thank you for the update, I am unblocking as I trust your message. — xaosflux Talk 13:31, 4 April 2016 (UTC)

Bot feedback idea

I just saw a bot revert an edit automatically and the edit summary was "Reverting possible vandalism by ## to version by ##. Report False Positive? Thanks,". And the core of it felt like "hey did I do the right thing? tell me? tell me?". The edit was very much justified but there's no way to give that feedback automatically. So I thought now that the "thanks" system is in place. Why not enable it so that one can thank a bot which provide automatic feedback for future automated reference or to strike pages of some internal checklist for manual review. A "no, bad bad bot" links will perhaps also be necessary to not bias the feedback data. Just an idea.. Bytesock (talk) 20:52, 11 April 2016 (UTC)

So, you're proposing an "anti-thanks" system that bots can then read via the API? Sounds like an great idea! Of course, we'd have to restrict it to bots (imagine a regular editor "anti-thanking" another editor in an edit war). APerson (talk!) 00:36, 12 April 2016 (UTC)
Can't an editor just revert a bot's edit? That should result in a notification, just as a Thank does. – Jonesey95 (talk) 00:48, 12 April 2016 (UTC)
This idea is absolutely redundant. The bot already has an interface you can go to where you input the code of the revision that it left in the edit summary. You can then tell the bot that it was an erroneous edit. That automatically improves the bot.—cyberpowerChat:Limited Access 16:40, 13 April 2016 (UTC)

Countdown to server switch

This is your reminder that all the wikis will be in read-only mode for about half an hour at 14:00 UTC (about 15 hours from now).  The announcements are at m:Tech/Server switch 2016 and on the blog. After we can edit again, I'll check WP:VPT for problem reports or advice. Don't forget that this will happen again on Thursday. Whatamidoing (WMF) (talk) 23:20, 18 April 2016 (UTC)

Where's the opt out button? *looks furiously*.. If only there was one. -- Cheers, Riley 23:23, 18 April 2016 (UTC)
I'll get the popcorn to watch it all go haywire. :p—cyberpowerChat:Online 7:26 pm, Today (UTC−4)

Public code

The mess above suggests to me that we need to change bot policy as follows: No bot will be approved unless its code is publicly posted, on-wiki. If the code is removed or not current, the bot must stop until its posted. We currently have a malfunctioning bot, and an editor who would be willing to fix the code, but not to rewrite the whole thing. Requiring public code would prevent this problem for future bots. Oiyarbepsy (talk) 05:02, 11 April 2016 (UTC)

What about AWB bots and other such things? I understand the sentiment here, but I'm not sure whether this requirement would result in less bot ops. Some people don't like to their code to be open source. ~ RobTalk 05:05, 11 April 2016 (UTC)
Right. It may be better to have the code only accessible by a select group of editors who are identified by the foundation (such as WP:OVERSIGHT currently is). That way, any code where its author has not authorized it to be "open source" can hold someone responsible if the code gets leaked ... with the exception of extreme cases where it is specifically has to be used by a new bot to replace their own? Steel1943 (talk) 05:20, 11 April 2016 (UTC)
AWB itself isn't a bot; it's a program that is under supervision of several active users. Individual AWB bots can post their AWB configuration files in an appropriate location - this would allow for any necessary debugging or cloning of the bot when necessary. עוד מישהו Od Mishehu 05:29, 11 April 2016 (UTC)

Now wait one minute? Isn't a huge part of the Wiki-mission to encourage open-source content? How are we encouraging open-source content by allowing close-source code to run our software? Oiyarbepsy (talk) 05:26, 11 April 2016 (UTC)

The encyclopedic content is open-source; some of the bots aren't - but their not part of Wikipedia, they only help maintain it. And many bots have BEANS issues, just as some of the edit filters are hidden from public view. עוד מישהו Od Mishehu 05:29, 11 April 2016 (UTC)
Right. Sometimes bot code has good reason to NOT be open source. For example, ProcseeBot's proxy-finding code could be used for malicious purposes on other sites. (This exact example used to be mentioned at Wikipedia:Creating_a_bot#Open-source_bots, until it was removed as part of a unilateral overhaul of the page. Additionally, Oiyarbepsy proposes to require the code to be posted "on-wiki", which many bot operators, even those that currently open-source their bot code elsewhere, would object to; doing so would automatically cause the code to be licensed under the CC-BY-SA 3.0 license, which might not be compatible with the license(s) of other code they may be using. Also, Creative Commons itself explicitly recommends against applying its licenses to software. jcgoble3 (talk) 05:32, 11 April 2016 (UTC)
I can say that I would not post the configuration files of my AWB bot on-wiki for exactly that reason. I have no objection to the regex I write being licensed under CC-BY-SA 3.0, but I don't care to monitor the licensing status of AWB itself. I much prefer a requirement of activity (discussed in the section above) over a requirement of open source. ~ RobTalk 05:45, 11 April 2016 (UTC)
I second Od Mishehu's and Jcgoble3's concerns. WP:BEANS, and CC recommends not using CC licenses for software. Also, if a bot uses code that licensed under a licence that is not compatible with CC-BY-SA 3.0, it cannot be posted on wiki. — JJMC89(T·C) 06:20, 11 April 2016 (UTC)
I myself am open source, but object to onwiki. If that were ever a requirement, I wouldn't be a botop. License issue aside, wikis, are no substitute for code repos, and the maintainability of updating individual files, for multi-file bots is ridiculously tedious, and would require a bot to do for you.—cyberpowerChat:Offline 07:04, 11 April 2016 (UTC)
  • What we are missing is why this is a headache - bots are supposed to help an editor do something boring and repetitive - but the worry is that other editors have come to depend on one editor (via their bot) to do something - if these tasks are so important we need to fund a community tool server to run these as a group - with open code. — xaosflux Talk 11:24, 11 April 2016 (UTC)
  • I would also support the public code initiative but oppose to publish the code on-wiki. To encourage this somebody couls create a template and tracking categories (Category:Bot with public source code, Category:Bot without public source code). Also {{bot}} could be extended to include a paremeter for the url of the source code. – T.seppelt (talk) 11:50, 11 April 2016 (UTC)
    Category:Wikipedia bots with source code published already exists. — Earwig talk 16:24, 11 April 2016 (UTC)
  • Uhh, publishing code on pages is ridiculous. People should be encouraged to use the proper solution, VCS, instead. If the ability to reuse bot code is important, I'd rather see Github's fork button than "now copypaste the content of these 12 pages". Max Semenik (talk) 15:16, 11 April 2016 (UTC)
    This is what I mean. Let the bot operator choose her/his preferred solution. A link to the repository (or in case to the wiki page) is enough. Tracking bots with public code could encourage this. -- T.seppelt (talk) 15:51, 11 April 2016 (UTC)
  • Absolutely not. Requiring public code is a horrible requirement. Editing Wikipedia is a donation of time and knowledge. We don't demand that Internet Explorer users only use an open source browser to contribute. We allow anonymous edits! We don't demand that photographers upload their entire portfolio. Bots are not part of the Wikipedia. If a bot owner wants to open source their code, that's up to them and we should encourage that. But never require it. Never. Freedom and openness demand that we allow choice. -- RM 16:16, 11 April 2016 (UTC)
    • I see no reason to require that bots be open source, even in cases where BEANS isn't an issue. Are we better off with a bot with complex amount of work behind, which does a very useful task, but the owner is complete;y unwilling to release the code for; or without it? I certainly think that with it is better. And while the Wiki content of the site is open source and free, I see no reason why all bots should be. I think that (except for BEANS cases) it should be encouraged, but never required. עוד מישהו Od Mishehu 18:09, 11 April 2016 (UTC)
Comment This seems like a mandate in search of a bot operator to mug. If you want to make progress on this, start by making it mandated in WP:BOTPOL in contravention of Authors of bot processes are encouraged, but not required, to publish the source code of their bot. (which you've seen several editors oppose) and requiring BAG to decline BRFAs if they don't meet the requirement. Several bots have a secret sauce or a blend of 11 herbs and spices that makes them go, in which case they should not be published. As to requiring it to be published on wiki, that is a unbelivably bad idea short of the foundation/labs providing a version control bridge that interfaces with normal version control software and the "approved" storage. Hasteur (talk) 20:08, 20 April 2016 (UTC)

ClueBot III still down

Before the servers get to move one last time for now, it seems to bother many that ClueBot III is down since last February. And there seems to be no discussion here, although it is said that Cobi had been doing his job to try to fix the problem somewhere else. Any news? 49.148.27.180 (talk) 14:00, 21 April 2016 (UTC)


On March 21, 2008 (yes that far back) and for a few days, User:CorenANIBot got into a odd task and copied it seemed the discussions from that day onto separate subpages. Some are just copies of the discussion, some have other people's edits there. One of those subpages is up for deletion at MFD here. -- Ricky81682 (talk) 20:12, 21 April 2016 (UTC)

Incorrect category creations

(crossposted from User talk:DumbBOT) This bot is incorrectly taking some categories by misposting some templates, see Category:Wikipedia files with no copyright tag as of 5 April 2016 and Category:Wikipedia files with no copyright tag as of 6 April 2016 for examples. It seems to be limited to the "files with no copyright tag" categories. It looks like the bot creator is no longer active but perhaps another editor can help resolve this problem. Liz Read! Talk! 15:18, 3 April 2016 (UTC)

Hopefully this isn't another Rotten Tomatoes (see above) situation where a botop has left but left their bots running. — xaosflux Talk 15:57, 3 April 2016 (UTC)
Template:Images with no copyright tag subcategory starter was deleted following Wikipedia:Templates for discussion/Log/2013 March 5. --Redrose64 (talk) 16:29, 3 April 2016 (UTC)
Based on that TfD and this AN discussion, this bot likely needs to be blocked until the botop returns. Other botops can adopt any useful tasks in the meantime/indefinitely. ~ RobTalk 16:42, 3 April 2016 (UTC)
I'm 100% in support of broken bots being blocked, however if the bad actions are significantly outpaced by good actions, a bit of time may be warranted for reply. How many bad edits are being made? — xaosflux Talk 00:02, 4 April 2016 (UTC)

One per day among 50 useful edits, some of which are important to the project. Maybe it would be beneficial for some bot-ops to start work to adopt the two remaining tasks (creation of daily subpages and removing protection templates from unprotected articles). In the meantime, an admin could preemptively salt Category:Wikipedia files with no copyright tag as of 8 April 2016 through Category:Wikipedia files with no copyright tag as of 30 April 2016 to buy us development time. ~ RobTalk 00:14, 4 April 2016 (UTC)

I'd rather speedy delete them then salt them. — xaosflux Talk 00:26, 4 April 2016 (UTC)

BTW, there is also some discussion of these issues at the bot's talk page. Speedy deleting isn't enough, as Dumbbot will live up to its name and recreate the categories later. So, any creation protection also needs to apply to categories that have already been speedy deleted, at least until the date on the category has passed. Oiyarbepsy (talk) 00:25, 11 April 2016 (UTC)

From User_talk:DumbBOT

@Tizio: WP:PUF has now been closed. Could you please have DumbBOT stop creating daily subcategories of Category:Wikipedia possibly unfree files? GeoffreyT2000 (talk) 23:33, 9 April 2016 (UTC)

Forgot a ping, so adding it now: @Nyttend:. Oiyarbepsy (talk) 22:34, 10 April 2016 (UTC)

  • I've just removed "Wikipedia files with no copyright tag as of DATE". Let's see whether this helps. Some time ago, I added a no-longer-used category to the title blacklist, because we'd deleted the category tree but couldn't otherwise stop DumbBOT from creating it; the bot's not an admin, so the blacklist works well. If removing the line from CatCreate works, we can just remove the PUF line as well; if not, we can just blacklist the titles. It's not like newbies are going to be doing good-faith creations of categories with these names, and the experienced user who's just out-of-date will know how to discover what's going on if he's blacklisted from creating such a category. Nyttend (talk) 22:45, 10 April 2016 (UTC)
  • Thanks for letting me know. I reverted the bot and protected the page (using TP because it's the lowest level of protection that the bot can't edit through), so let's see what happens next. Nyttend (talk) 00:03, 11 April 2016 (UTC)
  • Even worse, it re-created the bogus categories. [[:Category:Wikipedia files with no copyright tag as of 10 April 2016(->)]] as an example. Will nominate for speedy again, this time requesting creation protection. Oiyarbepsy (talk) 00:13, 11 April 2016 (UTC)
  • Self-reverted the protection and removal. The bot couldn't create the "no copyright tag" categories until a few weeks ago, when the blacklist entry was removed; I'm not clear why. Nyttend (talk) 00:37, 11 April 2016 (UTC)
  • It's on the blacklist talk page: "OK, I've removed this - we expect bots to edit constructively and not depend on the title blacklist, if it misbehaves it can be shut down.". That's why. I kinda agree with Xaosflux, a title blacklist is the lamest possible hack rather than a real fix. Oiyarbepsy (talk) 01:30, 11 April 2016 (UTC)
  • I'm ready to block this bot tomorrow if the operator continues to be absent - making project hacks around the malfunctioning bot is not our responsibility. Nyttend, I won't cry wheel if you revert that titleblack list - but lack of operator attention is not tolerable. -- xaosflux Talk 01:32, 11 April 2016 (UTC)
  • What's going to cause more of a problem: a bot creating a few extra categories, or not having a lot of these maintenance categories and other basic project pages? Until another operator's ready to have a bot perform the same tasks, or until DumbBOT starts doing things that are actual problems, its actions are doing more good than harm, and a block will do more harm than good. Those bad actions are, indeed, significantly outpaced by good actions, so unless you're going to create all those pages manually, a block will require a lot more work on someone else's part. At the moment, the project hack is the simplest way to prevent this problem without losing the benefits of the bot's far-more-numerous good edits, so it ought to be used until someone's gotten a replacement bot up and running. Once we have someone ready to take over DumbBOT's functions, block away, but as Cyberpower doesn't have the time to write it all anew and nobody else has chimed in at WP:BON (unless I missed something), we shouldn't do anything yet. Nyttend (talk) 01:42, 11 April 2016 (UTC)

I'll hold off on blocks while this discussion continues - but we have a core issue that needs to be resolved: bot operators are personally accountable for all actions made by their bots, and this operator is absent. -- xaosflux Talk 02:13, 11 April 2016 (UTC)
[End text copied from the bot's talk page]

  • @Xaosflux: I totally agree with the issue. I think we may need to propose some sort of new policy which requires that a bot's operator have a certain amount of activity within a certain amount of time (such as what happens with administrators after a certain amount of time), such as emailing a "bot-operator-verifier" team once a week or so to confirm that they are still watching Wikipedia in case their bot needs assistance. And maybe, if there is proof of their inactivity, then the bot would either be blocked, or the bot's code turned over to someone else to start a new bot, the latter only applying in cases where the bot being blocked would break more than it would help (as shown here). If we have to store the bot's code with a specific team, maybe they would need to have their identity confirmed (such as what is done with WP:OVERSIGHT) so there is accountability for seeing the code. Steel1943 (talk) 02:22, 11 April 2016 (UTC)
    • Once a week is way too frequent, but I do agree there should be requirements for operator activity. My recommendation would be one edit in the last six months, with a month of warning (by talk page and email) before a block. That's an intentionally low bar, but if an editor is paying enough attention to remember to make their one edit, then they'll probably respond if prodded on talk/email to do so. ~ RobTalk 02:25, 11 April 2016 (UTC)
      • I don't think we need them to proactively prove they are live, so long as they are required to re actively reply to requests within a reasonable period (up to a week from either their or their bot's last edit?) -- xaosflux Talk 02:29, 11 April 2016 (UTC)
        • @Xaosflux: I think the issue is that we're always behind if we just react to inactive operators. If we have a month where we know they have to respond or be blocked, other operators can start preparing to take over their bot's duties. It's a significant advantage to be able to do that in an environment where the bot isn't already broken and actively editing in error (or worse, not working entirely). ~ RobTalk 02:33, 11 April 2016 (UTC)
        • (edit conflict) @Xaosflux: If that idea means that it requires that a week pass after an edit from the bot and its operator, then this would not apply to DumbBOT's situation since the bot is still actively editing. I would just apply this time requirement to the operator since bots will run or not (especially if it is blocked) regardless of operator participation. (And yes BU Rob13, a week is probably too frequent for my idea; I, was more or less, trying to start a conversation here that may result in a new, and seemingly necessary, new bot/operator accountability policy.) Steel1943 (talk) 02:36, 11 April 2016 (UTC)
    • [double edit conflict] Take that to WP:VP/Pr, and I'll heartily support the general idea, although I agree with BU Rob13 that activity requirements should be less. The only reason I'm opposed to blocking DumbBOT right now is that it's doing a ton of good work, without interruption and without error. It's the perfect bot task, running basically as an add-on to the Mediawiki software by creating stuff that would be done automatically by the site software if that were possible. And yes, the operator isn't responsive, and that's a problem that the bot policy ought to cover (I didn't see it in a quick run-through of the policy), but this is a perfect WP:IAR situation, in which we don't stop the bot until its functions have been redundified (new word?) by someone else. Nyttend (talk) 02:31, 11 April 2016 (UTC)

Sorry for the inconvenience. I removed that page from the category creation bot function (yes, the list is merely a report of what the bot is doing). If someone wants to take over the whole function I will shut it down. Otherwise, in case of problem please email me, because I really cannot guarantee logging in often. I added such a notice to the bot's page. Not being an active user any longer, I do not think I should say anything on the policy about bots, just let me know if/when I have to shut it down altogether.Tizio 13:25, 11 April 2016 (UTC)

(Note: Cyberpower678, see above.) Steel1943 (talk) 15:05, 11 April 2016 (UTC)
A link to the source could would be nice.--cyberpowerChat:Online 19:47, 11 April 2016 (UTC)

Test run - the bot created two categories: Category:Wikipedia files with no non-free use rationale as of 11 April 2016 and Category:Wikipedia files needing editor assistance at upload as of 11 April 2016. I believe these are ok, since categories for other days exist and are not empty. A wrong category may still be created this night (the bot runs on two computer, at the moment I can't access the secondary one), but should be the last one. Please let me know if there are some other category that is incorrectly created.Tizio 18:39, 12 April 2016 (UTC)

Hazard-Bot creations

For some reason, User:Hazard-Bot created Category:Possibly unfree files from 2016 April 21 and Category:Possibly unfree files from 2016 April 20. Hazard-Bot began this "task" on the 17th and the prior ones have all be deleted. I've already informed User talk:Hazard-SJ but (a) when did we authorize two bots to do the same task and (b) is this another bots that will take on this task if we delete these? -- Ricky81682 (talk) 20:11, 20 April 2016 (UTC)

We certainly can have multiple bots that perform the same tasks - however they must cooperate and not duplicate work. — xaosflux Talk 20:29, 20 April 2016 (UTC)
Sorry, that wasn't my question really. I guess the question is, if there's multiple bots doing this category creation, was each one approved for that and if so, is there a central place to actually say stop to every one of them? Or is this just one by one us figuring out which bot is doing the same task? -- Ricky81682 (talk) 20:42, 20 April 2016 (UTC)
We don't have an index by "task" - you could go try going through Category:Approved_Wikipedia_bot_requests_for_approval for the more recent ones. — xaosflux Talk 21:36, 20 April 2016 (UTC)
@Ricky81682: Thanks for the notification! I wasn't aware the PUF process had ended, and my bot has actually been doing that task (or at least, while it hadn't been beaten to it by another bot) since September 2013. I have, however, stopped the PUF category creation in this commit.  Hazard SJ  21:09, 23 April 2016 (UTC)
@Hazard-SJ: Great. Oddly enough, someone had already started a number of discussions on the 22nd so I moved those to FFD for now. -- Ricky81682 (talk) 22:56, 23 April 2016 (UTC)

Do global bots need local authorization?

Question's right up top. --QEDK (T C) 17:10, 2 May 2016 (UTC)

WP:GLOBALBOTS are only allowed to update interwiki links. –xenotalk 17:27, 2 May 2016 (UTC)
What about fixing double redirs? --QEDK (T C) 18:23, 2 May 2016 (UTC)
Would require local approval, I'd think. –xenotalk 18:31, 2 May 2016 (UTC)
Concur here - if doing anything else attempt to warn the operator, if not responsive or being disruptive we can block. — xaosflux Talk 18:50, 2 May 2016 (UTC)
Bot in question: Invadibot, Bot op notified. --QEDK (T C) 19:01, 2 May 2016 (UTC)
Thank you, I left a note on the op's home wiki as well es:Special:Diff/90828931. — xaosflux Talk 20:35, 2 May 2016 (UTC)
Red X User blocked pending operator response as edit have continued after messaging. — xaosflux Talk 20:42, 2 May 2016 (UTC)
 Resolved Bot operator has made adjustment, block has been removed. — xaosflux Talk 02:12, 3 May 2016 (UTC)

Legobot error: Good article nominations

As documented at Wikipedia talk:Good article nominations#Legobot error?, we've positively identified a bug in the Legobot code that is supposed to pass good articles. The bug is: If a reviewer passes an article as GA and places the passed {{GA}} template immediately above an earlier placed {{failedGA}} template from an earlier review, the bot will become confused by the presence of both templates and will fai33l the GA rather than pass it. It happened again today at Talk:Gene Roddenberry. It happens nearly every day. When it happens, the successful nominator receives notice that the article failed instead of passed, and the GA icon is not placed on the article; it must be manually placed. A possibly separate bug: Sometimes the GANotice template is not placed on the nominators talk page at all. Pinging Legoktm; can you please help us. Side note: We're also wondering if you can help us add a few subtopics as documented at Wikipedia talk:Good article nominations#RfC: Subcategories on GAN page. Thank-you. Prhartcom (talk) 22:28, 15 March 2016 (UTC)

  • The errors continue to occur, causing both incorrect edit summaries and incorrect posts to user talk pages. Is this kind of delay in responding to a report here typical? Legoktm (Legobot), it would help a great deal if we knew about how long a fix of this bug is likely to take. Thank you for your attention to this matter. BlueMoonset (talk) 01:48, 31 March 2016 (UTC)
  • Now up to over seven weeks here without a response. We continue to need what Legobot does for the GA process, quite desperately, and are most grateful for it—I still remember that awful period after GAbot went down before Legoktm incorporated the code into Legobot—but none of us realized that we'd be frozen there without any possibility of future improvement or bug correction. Do any of the people who frequent this noticeboard have any suggestion as to how we might proceed? BlueMoonset (talk) 05:41, 6 May 2016 (UTC)
  • I'd recommend WP:Bot requests with a request to take over the task. The Bot owners' noticeboard isn't really well-suited to getting a fix for a bot. The bug isn't so bad to warrant blocking the bot, and communication doesn't appear to be working, so I don't know how much else can be done at this venue. ~ RobTalk 07:18, 6 May 2016 (UTC)

Very minor: topicon updates

Hi, I updated Template:Bot topicon slightly, to allow specification of owner, icon customization, and custom id (for sorting). I also made Template:Bot operator topicon with a new LGPL image, File:BotOp-logo.svg which also allows image/id customization and bot specification. — Andy W. (talk ·ctb) 16:10, 9 May 2016 (UTC)

"Welcome" message

Please report here any unusual errors if your bot runs in to the "welcome" message trying to make an edit. Trying to track any of these down. — xaosflux Talk 20:27, 19 May 2016 (UTC)

Need developers

Ladies/Gents - I know you're sitting at home right now, probing Wikipedia, and looking for something to soften that boredom. You could work on an article about the 1968 Philadelphia Eagles season, but the tediousness of finishing the whole 1960s collection is just dragging on and on. Let me offer you a new opportunity! The Unblock Ticket Request System is looking to get some additional developers on board. Developers do not have to be administrators, but administrator tools will be required for SSH access and access to the live instance of the tool. Non-administrators can still develop with Github and will have access to UTRS-Alpha and UTRS-Beta. Specifically, we're looking for two associate core developers, a senior user experience developer, and a user interface developer. Specific requirements are listed on the link above, but we're looking for experience with HTML5, PHP, jQuery, CSS3, and bootstrap. Additional cross-platform experience, team management, object oriented design, and secure systems development experience is a huge bonus. Please send an application to utrs dash developers AT googlegroups dot com. We have big ideas and big plans for a v2.0 of the system and we're looking for some folks looking to do exciting work.--v/r - TP 03:14, 23 May 2016 (UTC)

Notice of speedy deletion change

It has been agreed that categories must be empty for seven days (not four) before being speedy deleted. If you have any bots or other automated processes based on this, place make the appropriate changes. Oiyarbepsy (talk) 04:51, 24 May 2016 (UTC)

Insecure (non-HTTPS) API Requests to become unsupported starting 2016-06-12

Please ensure your bots are using HTTPS: links to access APIs of Wikimedia-hosted production wikis! Wikitech-l announcement email with deeper details (also cross-posted to mediawiki-api and mediawiki-api-announce): http://lists.wikimedia.org/pipermail/wikitech-l/2016-May/085618.html --BBlack (WMF) (talk) 22:43, 13 May 2016 (UTC)

Looks like this will be the death of User:Bibcode bot, unless someone is willing to help me convert the code. Converting my bot from pywikipedia to something modern seems to be well beyond my skills. Headbomb {talk / Special:Contributions/Headbomb / physics / books} 12:17, 20 May 2016 (UTC)
Do you have a pointer to a public source repo for the Bibcode bot? AIUI, pywikipedia was replaced by (or renamed to?) Pywikibot, which does do HTTPS. --BBlack (WMF) (talk) 19:32, 20 May 2016 (UTC)
Spent a few hours trying to setup a repo, and it seems downright impossible to do so, so here's zip files instead. Here are the only files (I think) I use on top of some old pywikipedia install (http://www.file dropper.com/bibcodebotminimal), but if that doesn't work here's my complete pywikipedia folder (http://www.file dropper.com/bibcodebotfull). Remove the space between file and dropper. Headbomb {talk / contribs / physics / books} 21:47, 20 May 2016 (UTC)
@Headbomb: If you dont get help before June 1, shoot me and email and I'll do the conversion for you. John Vandenberg (chat) 01:34, 21 May 2016 (UTC)
@John Vandenberg: Much appreciated! Headbomb {talk / contribs / physics / books} 02:05, 21 May 2016 (UTC)

All AWB API requests are HTTPS unless forced otherwise. No AWB bots will break. -- Magioladitis (talk) 12:34, 20 May 2016 (UTC)

I'm worried about this list of users, whose bots or scripts are actively using http:// and whose owners look hard to contact:

Account Contribs Probable owner
User:Nuno Tavares Special:CentralAuth/Nuno Tavares User talk:Nuno Tavares
User:FacebookBot Special:CentralAuth/FacebookBot User talk:Facebookesun?
User:AdminStatsBot Special:CentralAuth/AdminStatsBot User talk:JamesR
User:BracketBot Special:CentralAuth/BracketBot User talk:A930913
User:DamianZaremba_Scripts Special:CentralAuth/DamianZaremba_Scripts User talk:DamianZaremba
User:DumbBOT Special:CentralAuth/DumbBOT User talk:Tizio
User:LaSabiduria Special:CentralAuth/LaSabiduria User talk:LaSabiduria
User:MadmanBot Special:CentralAuth/MadmanBot User talk:Madman
User:MilliMar Special:CentralAuth/MilliMar User talk:MilliMar
User:Ananthanns Special:CentralAuth/Ananthanns User talk:Ananthanns

I'm going to leave messages on the named talk pages, but I'm not expecting much of a response. Several of these editors haven't edited for months. If anyone knows anything about these bots, please {{ping}} me. Whatamidoing (WMF) (talk) 18:20, 20 May 2016 (UTC)

@Whatamidoing (WMF): User:FacebookBot says it is operated "by the foundation" - is there actually a staff member assigned to this still? — xaosflux Talk 18:50, 20 May 2016 (UTC)
The Facebook one? I'm still searching. Certainly Philippe (who needs to login and update his user page again) is no longer responsible for that one. Whatamidoing (WMF) (talk) 19:05, 20 May 2016 (UTC)
Thank you (I forgot to copypaste that userid). — xaosflux Talk 19:29, 20 May 2016 (UTC)
@Whatamidoing (WMF): "Contacts: For more information, contact Facebookesun (talk · contribs) or Philippe (WMF) (talk · contribs)" I would recommend contacting Facebooksun, a Facebook employee that is responsible for the bot, IIRC philippe was only a liaison with Facebook and had nothing to do with the actual bot. Peachey88 (T · C) 10:33, 24 May 2016 (UTC)
Thanks. I left a message. However, the user hasn't logged in for more than five years, which is rather longer than the median employee tenure in the tech industry, so I'm not very hopeful. James A is passing the message along through other channels. Whatamidoing (WMF) (talk) 15:15, 24 May 2016 (UTC)

If anyone's getting a "500 Can't verify SSL peers without knowing which Certificate Authorities to trust" error, then there's a possible solution at WP:VPT. Whatamidoing (WMF) (talk) 23:03, 20 May 2016 (UTC)

https api change

@Whatamidoing (WMF): RE: I don't bot op as much as I used to. Can you please tell me (or direct me to) what requests are coming through regular API and I will try to make improvements on it. Or better yet guide me as to where I can self service my requests. Hasteur (talk) 18:04, 20 May 2016 (UTC)

HasteurBot turned up in a list of bots that made between 100 and 999 requests over http:// during one week (the week ending 13 May). Given the bot's activity, I'm going to guess that it's all of them. But that's just a guess, so let me track down someone who can give you more information. Whatamidoing (WMF) (talk) 18:54, 20 May 2016 (UTC)
HasteurBot's requests in the past 24H all came from two user agents: g13bot_tools-g13_interested_notify.py/r-1 (unknown) Pywikipediabot/1.0 and g13bot_tools-g13_nudge_bot.py/r-1 (unknown) Pywikipediabot/1.0, and seem to originate in toollabs exec nodes. All the requests were for enwiki APIs. --BBlack (WMF) (talk) 19:18, 20 May 2016 (UTC)
Yep, that's what I was afraid of. Time to retire the old pywikibot-compat code and get these running on the core. Hasteur (talk) 19:29, 20 May 2016 (UTC)
@Hasteur:, yell out if you need help, either due to difficulty or you just dont have time. Pywikibot 2.0 is the most stable, but it uses httplib2, whereas 3.0-dev is a little more raw and uses requests. Either come on #pywikibot connect, or create a task in Phabricator if you want Pywikibot devs to send you pull requests. John Vandenberg (chat) 01:32, 21 May 2016 (UTC)
Admins (Whatamidoing (WMF)BBlack (WMF)John Vanderberg) I think I've corrected the G13bot (several tasks authorized for HasteurBot) API requests. I request a API pull check in 24 hours starting at this timestamp to verify I have this resolved. Also is DRN clerk bot also on the piggy list? I'm also the lever puller for that and if I need to yank some time to address that, I need to do it sooner rather than later. Hasteur (talk) 22:28, 23 May 2016 (UTC)
My list only included accounts that made 100+ edits via http:// during a one-week period. DRN clerk bot wasn't in my list, but that doesn't mean that it couldn't have made ≤99 insecure edits. Whatamidoing (WMF) (talk) 23:31, 23 May 2016 (UTC)
@Hasteur: If your bot uses a distinct User-agent header, which it looks like it does, you can check it yourself using Special:ApiFeatureUsage. But to save the effort at the moment, the last log entry for this issue that I see for User:HasteurBot is at 2016-05-22T03:00:31Z. I see no hits for User:DRN clerk bot in the past week (although it is using either the empty string or 0 somewhere instead of "now" to represent the current time). BJorsch (WMF) (talk) 21:29, 24 May 2016 (UTC)
Doesn't work for me. I essentially copied the UA being used by Cyberbot, into that form and nothing turned up.—cyberpowerChat:Online 21:48, 24 May 2016 (UTC)
User:Cyberbot doesn't seem to do much of anything, but I see plenty of stuff for the Peachy agent used by Cyberbot I and a few for the other one too. BJorsch (WMF) (talk) 22:37, 24 May 2016 (UTC)

Aw crap. What a pain in the ass. Hawkeye7 (talk) 13:26, 24 May 2016 (UTC)

Dualling bots

A relatively small matter, but unfixed and replicated across a lot of articles it could be a much larger one. At Zahi Hawass three edits in a row in three days were by bots. Yobot made a change. GreenC bot undid the change (and did something else), Yobot redid the change. Clearly not both reading from the same style book. Neither actually fixed the problem which is a comma next to a full stop, which maybe means this will not happen elsewhere. But when doing general fixes they should be following exactly the same rules otherwise this could happen again in other ways.--JohnBlackburnewordsdeeds 08:32, 7 June 2016 (UTC)

User:GreenC bot appears to be mostly operating correctly (Special:Diff/723950617 copyedit failed to improve the page due to the original authors invalid syntax). User:Green Cardamom, please check. — xaosflux Talk 10:54, 7 June 2016 (UTC)

This seems to be more of a WP:COSMETICBOT problem with Yobot.

@Magioladitis: Please review, and identify what is going on with these edits (i.e. why they are not COSMETICBOT violations):
  1. Special:Diff/723882099
  2. Special:Diff/724112506
  3. Special:Diff/724112499
  4. Special:Diff/724112380
  5. Special:Diff/724112344
xaosflux Talk 10:54, 7 June 2016 (UTC)

JohnBlackburne and Xaosflux My specific bot's edits are part of CHECKWIKI (error number 61 Reference before punctuation) and are done per Manual of Style. More specifically per WP:REFPUNCT. GreenC bot has to take care of punctuation to be BEFORE references. Take not that Yobot's edit changes the visual outcome so it si not cosmetic. Correct punctuation was/is a criterion of creating good articles. Error 61 is fixed by Yobot since June 2010. -- Magioladitis (talk) 11:04, 7 June 2016 (UTC)

Aha. JohnBlackburne beats us all :) There was a punctuation duplication causing the error. JohnBlackburne fixed it. -- Magioladitis (talk) 11:05, 7 June 2016 (UTC)

GreenC bot doesn't mess with punctuation. What happened is GreenC bot inadvertently reverted Yobot's edit. GrenC does so much IO it takes a long time, so it runs off-line generating a diff. When that diff is later uploaded (via AWB) it checks the article size to see if there were any changes and re-runs real-time if there are. Since the changes made by Yobot did not change the article size that is the problem. So what I need is a better way to check for a diffs than article size (duh). That will be no problem, sorry for the revert. -- GreenC 12:49, 7 June 2016 (UTC)

Ah, I see. So my concern that the bots had different ideas of the 'correct' style was unfounded. But yes, assuming an article is unchanged because its size is unchanged is going to get it wrong often; not only bots but editors can make edits that purely rearrange things, or change other things that make no difference to the size.--JohnBlackburnewordsdeeds 12:26, 8 June 2016 (UTC)
Yes it was an oversight now fixed. It wouldn't have happened that often because it required two conditions, the change to result in a 0-length difference and the change to occur within the time window between when the bot retrieves the article processes it and uploads the diff. -- GreenC 12:45, 8 June 2016 (UTC)

I am posting this issue here because I am uninformed of our bot policy and would like more informed editors/admins to take a look at this. I have an outstanding request that I am working with @BU Rob13: at Wikipedia:Bot requests ({{WikiProject Green Bay Packers}} Banner is the relevant section). @Magioladitis: (i.e. User:Yobot) seems to have hijacked our conversation and begun editing without any approval from me, Bu_Rob13 or WP:BAG. Please note that the bot's only contributions so far have been cosmetic, such as this. Check the bots contribution history and see how almost (if not all) of the edits so far have been cosmetic. Please note the conversation at User talk:Yobot. If anyone could advise on the proper actions, I would appreciate it. « Gonzo fan2007 (talk) @ 18:03, 8 June 2016 (UTC)

I'm really not interested it getting involved in this, but I would like to note that I think Yobot has a very old approval from BAG (pre-2010) to do literally anything involving project tagging. Changing parameters, tagging articles, etc. This is presumably the approval he was using to make his edits. I'm not sure why the cosmetic edits were made, of course. Such an approval would almost certainly not be granted today. ~ RobTalk 18:21, 8 June 2016 (UTC)

BU Rob13 I am working on providing a custom module for general use instead of F&R rules because till now you use the latter and the result was duplicated parameters. Anyway, no big harm. I had no indention to hijack the discussion. I left my comment on the assessment of namespaces already in the BOTREQ. My apologies. -- Magioladitis (talk) 18:26, 8 June 2016 (UTC)

As I said on your bot's talk page, BU_Rob13 and I were still discussing how we wanted the articles tagged. He had open questions that were not answered by me. I don't care what you were working on, why would you have your bot start editing articles without asking either of us? Why is the bot performing cosmetic edits on talk pages, which is not per policy? Your edit summary invoked the Wikipedia:WikiProject Green Bay Packers name, which made it appear that it was sanctioned by the group, when it was not. Please explain. « Gonzo fan2007 (talk) @ 18:31, 8 June 2016 (UTC)

Let's don't have this discussion in many places. I can't reply in all places. All tagging bots I know till now use AWB general fixes and some custom module in addition to F&R rules. In the past some bad settings had caused broken templates/broken parameters etc. So, I always try to keep the custom modules others use up-to-date. I usuallly do this before other bots are about to perform a bot run. IT is the same reason I commented in the BOTREQ about the auto-class set for other namespaces. I apologise if this seemed as hijacking. I'll leave the discussion to you. -- Magioladitis (talk) 18:39, 8 June 2016 (UTC)

Request for Bot Approvals Group notification

As is required by the BAG membership procedure I am placing this notification at WP:AN, WP:VPM, WT:BOT, and WP:BON. I am requesting to join the Bot Approvals Group and my request can be found here: Wikipedia:Bot Approvals Group/nominations/HighInBC. HighInBC Need help? {{ping|HighInBC}} 20:24, 2 July 2016 (UTC)

BOT malfunctioning?

A bot has tagged an in-use sound file as an "orphaned image"[3] when it is in use at Antidisestablishmentarianism (word)) providing the US pronunciation. The file is also the target of a redirect. Is there a general problem? Last time there was large scale error-prone orphan file tagging it led to files being deleted as "unused, no foreseeable use" even when they were in use. Thincat (talk) 15:19, 4 July 2016 (UTC)

At first I thought it was because it was using a redirect on the page but after changing it to point directly to the file it is still not showing up in the File usage section which is what I assume the bot looks at.
It may be a side effect of how the file is being included on the page through the {{IPAc-en}} audio parameter rather than a conventional link.
I don't think the root of the problem is with the bot. As a short term fix I have reverted the bots orphan tagging of the file, I assume the bot will not edit war with me. HighInBC Need help? {{ping|HighInBC}} 15:32, 4 July 2016 (UTC)
@Fastily: Pinging bot operator. HighInBC Need help? {{ping|HighInBC}} 15:42, 4 July 2016 (UTC)
From memory (and this is way outside my scope) "file usage" has known problems and certainly is unreliable for determining whether a file is being used. "What links here?" may be better. Bots should not be run which rely on "file usage". Thincat (talk) 15:41, 4 July 2016 (UTC)
Yes, this sort of thing has happened before, back in 2012.[4][5] Thincat (talk) 16:10, 4 July 2016 (UTC)
The file isn't "used" in the sense that MediaWiki means when it considers a file to be used, i.e. being embedded in a page. Instead, the file is merely "linked". The bot that's doing the tagging should probably consider links sufficient to not be tagging it as orphaned, however. Anomie 21:14, 4 July 2016 (UTC)
I have clarified the wording of {{Orphan image}} to better explain its use in assisting maintenance/compilation of site statistics, and the particular type of links the bot looks for. Furthermore, I have implemented exclusion compliance in FastilyBot's task 10 (sorry I forgot :o), so that users are free to opt out for files they have uploaded. I hope this addresses Thincat's and Adam's concerns :) -FASTILY 21:19, 4 July 2016 (UTC)
No, my concerns are not addressed. Audio files linked to from articles should not be tagged as "orphan image". It is irrelevant what MediaWiki means by "used" – such files are being properly invoked by articles and are not orphaned (nor are they images). I do not have the ability to tag all audio files to exclude this bot, nor should editors be expected to do this. Similarly, images linked to should not be tagged as orphaned just because they are not being displayed. A linkage from an article is, I think, unusual, but it is common to link to a file from another file or from a talk page. I think in addition if an article invokes an image via a redirect, the target image is not "used". See WP:Village pump (technical)/Archive 127#File usage not shown. My concern is made greater because although {{Orphan image}} says it is not a deletion tag, it is used by some editors to submit files for deletion in a mechanistic manner. See WP:Files for deletion/2011 December 29#File:Melatonin-pronunciation.ogg and the following erroneous nominations. When files have been wrongly tagged with this template, some editors do not check whether they are in use before submitting them for deletion. It is important that files are not wrongly tagged as orphaned. Thincat (talk) 05:30, 5 July 2016 (UTC)
As per my previous reply to you, {{Orphan image}} is a tracking template that assists with maintenance and compilation of site statistics. Its intended use is for files without fileusage, which I agree, is different from linkshere. It is, however, by no means the deletion tag you're claiming it to be. I am not responsible for the way other editors use and interpret datasets I build; if other users are behaving disruptively, then this is a matter for ANI, which does not concern me or this noticeboard. -FASTILY 07:49, 5 July 2016 (UTC)
I note that, until your recent edits, the template was pretty clearly intended to track files for review for WP:NOTIMAGE concerns. Your edits have transformed it into a wishy-washy template that tracks things for no discernible purpose. You might also consider renaming it and its category to something closer to your intended purpose for it, e.g. "files not embedded in any pages", since "orphan" implies "not used" while "used" can include use via link rather than only via embedding. Anomie 15:50, 5 July 2016 (UTC)
There's an open TFD for {{Orphan image}} you might be interested in. I think your suggestion has merit, but given how there's a number of conflicting opinions regarding the future of the tag, I am hesitant to make any (more) drastic changes without some sort of community consensus -FASTILY 01:37, 6 July 2016 (UTC)

Template updater

Not sure if this is the right place but here goes. On another wiki, I'd like to use a bot to update and standardize the name of fields on a particular template. Since I'm still learning how to operate bots, is there a bot whose source code I could use/adapt to perform such a task? Abjiklɐm (tɐlk) 13:26, 5 July 2016 (UTC)

Have you tried the AutoWikiBrowser? The find and replace functionality on AWB can get this done without having to write any code. Ganeshk (talk) 20:48, 5 July 2016 (UTC)
I've used AWB on Wikipedia, but does it work with any wiki including those outside the Wikimedia projects? Abjiklɐm (tɐlk) 02:42, 6 July 2016 (UTC)
Yes. Σσς(Sigma) 06:23, 6 July 2016 (UTC)
Options > Preferences > Site, you can change the wiki there. Ganeshk (talk) 13:54, 6 July 2016 (UTC)

Unflagged bot XLinkBot

XLinkBot acts like a bot but does not actually have the "bot" flag. A bureaucrat should replace the "autopatrolled" and "extended confirmed user" rights of that user with the "bot" flag while leaving the "pending changes reviewer" right on. GeoffreyT2000 (talk) 19:09, 12 July 2016 (UTC)

@Beetstra:, @Versageek: - is there a special reason your bot needed to not be flagged? — xaosflux Talk 19:14, 12 July 2016 (UTC)
That's by design, I believe. Antivandal or antispam bots don't get these flags so that their edits - which by default may be inexact and error-prone - can be tracked more easily.Jo-Jo Eumerus (talk, contributions) 19:26, 12 July 2016 (UTC)
User:XLinkBot/FAQ #Why is XLinkBot not in the bot group and thus makes edits without a b tag? - NQ (talk) 19:28, 12 July 2016 (UTC)
Thank you, will let the operators have a chance to decide if this is still needed, other antivandalism bots such as User:ClueBot NG are now flagged (if I recall correctly it used to be harder to "showbots" in recent changes). — xaosflux Talk 19:31, 12 July 2016 (UTC)
Though it appears that is also a "technicality" - Cluebot NG has a bot flag, but doesn't assert the bot flag on edits, so it still appears in recent changes; while it may be better to give xlinkbot the bot flag for other reasons, it may require some programming updates to continue to appear. @GeoffreyT2000:, I suggest we move this to WP:BOWN for continued discussion. — xaosflux Talk 04:20, 13 July 2016 (UTC)
If this is practice, the bot policy should be updated to reflect that. If this is outdated, all of these bots should be flagged. Kharkiv07 (T) 23:30, 12 July 2016 (UTC)

(edit conflict)It has been a specific choice not to flag XLinkBot, and I think it is overly bureaucratic (no pun intended) to enforce that all bots should be that way. Some bots, like the antivandalism bots, do edits that do need constant scrutiny (to quick react on bad reverts, but also to catch the vandals/spammer - if RC patrollers see the bot revert a really bad link, they might get the editor blocked early). It is pertinent that bots that do 'gnoming' type of work (maintenance, wikidata moves) do not show up in RC feeds (they quickly flood feeds, they can even flood watchlists), but antivandalism bots should standard show up in the feeds.

I see that for ClueBot NG flagging has not been discussed, they have just been flagged after the approval (Wikipedia:Bots/Requests_for_approval/ClueBot_NG; approved on December 3, 2010, flagged on the same day). For XLinkBot this was discussed in the original request, see Wikipedia:Bots/Requests_for_approval/SquelchBot#Big_BAG_Section. Note that the bot's edits are 'manually' flagged in the edit-summary.

We could consider to give the bot the flag, and then de-flag the edits, but I don't see the advantage of that (it will just give me some programming work, and overall nothing changes ..). --Dirk Beetstra T C 05:12, 13 July 2016 (UTC)

@Beetstra: Thank you for the reply - I wouldn't say this is "urgent"; @GeoffreyT2000: - as stated above, and with Cluebot NG, the "bot flag" is not asserted by these bots when reverting - so it will not have any impact on watchlist/recent changes feed to fix it (if you want to argue that these bots should assert bot flag for these kinds of edits, that is a bigger discussion). — xaosflux Talk 11:45, 13 July 2016 (UTC)
Perhaps some of the other operators that use perl bots can chime in the assertion mechanism? — xaosflux Talk 11:45, 13 July 2016 (UTC)
It used to be that being flagged as 'bot' forced all the edits to be flagged as 'bot'. That hasn't been the case since the API action=edit module was created, though. Now, just don't specify the 'bot' parameter to action=edit and the edit won't be flagged. Anomie 00:45, 14 July 2016 (UTC)
Dirk Beetstra, are you ok to have us TRY to flag xlink bot, then make sure it is still not showing up with +b ? — xaosflux Talk 01:49, 14 July 2016 (UTC)
@Xaosflux: Let me first test it the other way around, so I know I can just turn it off if needed. --Dirk Beetstra T C 06:01, 14 July 2016 (UTC)
Thanks, please ping us back here - I can always flag you, make sure it is behaving as expected, and remove if it does not. Keep up the good work with your bot though! — xaosflux Talk 13:35, 14 July 2016 (UTC)

Last chance to fix these bots

Okay, here's the list of bots and script-users that are going to start breaking on Sunday, 12 June:

New – these owners haven't been contacted yet:

Previously known – these owners have been contacted, but haven't fixed the bots (list generated ~30 hours ago):

If you know how to reach these users, please help us contact them. I know that people are working on a couple of these, but I haven't heard from most of the affected bot owners. In about 72 hours, Ops is going to start refusing 10% of the http:// requests; a month from now, they'll break completely. You can reach me on wiki or via Special:EmailUser if something's sensitive. User:BBlack (WMF) and his team take questions on IRC, too.

Thanks for your help, Whatamidoing (WMF) (talk) 18:13, 9 June 2016 (UTC)

@Whatamidoing (WMF): what am I doing in that list? - DVdm (talk) 18:18, 9 June 2016 (UTC)
@WhatamIdoing: my semi-private Huggle 2 (User talk:DVdm#HG) perhaps? That would make sense. - DVdm (talk) 18:30, 9 June 2016 (UTC)
I made some updates to the list above and re(notified) operators on the top of the list. — xaosflux Talk 19:05, 9 June 2016 (UTC)
I am neither a bot owner nor a bot operator. --Nevit (talk) 19:50, 9 June 2016 (UTC)
Hi, my account (Der.Traeumer) is wrong in this list. I dont use a bot.--Traeumer (talk) 19:56, 9 June 2016 (UTC)
@Der.Traeumer: indeed not, but you are also using one of my Huggle 2 hacks, which still uses http. - DVdm (talk) 19:58, 9 June 2016 (UTC)
Maybe, but i only use it in dewiki, not here.--Traeumer (talk) 20:04, 9 June 2016 (UTC)
I expect this is a WMF wide break, not an enwiki one - I would expect you will break anywhere. — xaosflux Talk 20:44, 9 June 2016 (UTC)
By the way, I tried to force Huggle 2 to use https once. That did not work out well, so I gave it up. Alas. - DVdm (talk) 20:48, 9 June 2016 (UTC)

Xaosflux is correct: This is a global list, not solely for enwiki. The API's global, and the impending change affects all 800+ wikis – not just English-language ones and not just Wikipedias.

Arbitrary break 1

This is a list of "bots and script-users", regardless of whether the account has a bot flag. If you are using a custom AWB, Huggle, or similar script, then that is probably the cause. For that matter, any method of POSTing the API over http:// (e.g., to make null edits with a script, which results in some of the "zero contributions" accounts here) will put you on this list, because what you're doing will stop working soon. Whatamidoing (WMF) (talk) 20:56, 9 June 2016 (UTC)

Indeed. Thanks for the heads-up. As soon as my H2 dies, I'll try to have a look at the https thing again. If it still fails, I guess I'll have to go to Huggle 3—or stop unvandalising altogether. Bummer! - DVdm (talk) 21:05, 9 June 2016 (UTC)
Hi User:Whatamidoing (WMF). I see my bot, User:LivingBot, is listed. I thought I updated all of its code a couple of weeks back. What's the best way of narrowing down which of its scripts is still using an insecure connection? More generally, I suggest emailing operators (where operators have verified emails, naturally). - Jarry1250 [Vacation needed] 21:37, 9 June 2016 (UTC)
I'll request user agent information for you. Maybe that will help. Whatamidoing (WMF) (talk) 22:44, 9 June 2016 (UTC)

@Bgwhite: are you aware that your bot is listed here? -- Magioladitis (talk) 22:32, 9 June 2016 (UTC)

No, but Whatamidoing mentioned the problem earlier. I run AWB, WPCleaner and a pywikipedia script. All these have been "fixed" by their maintainer. I'm not sure what is causing the problem or how to narrow things down. Bgwhite (talk) 22:39, 9 June 2016 (UTC)
The long list above may be aggregated data from before the first post by BBlack on 13 May, maybe. — Andy W. (talk ·ctb) 23:03, 9 June 2016 (UTC)
@Bgwhite: The most-recent insecure-access log entry for account BG19bot was at 2016-06-09T08:27:28.000Z, with the User-Agent scripts-header_levels_labs.py/r11775 Pywikipediabot/1.0. The access wasn't from a labs IP address, it came directly from a residential ISP service. --BBlack (WMF) (talk) 01:05, 10 June 2016 (UTC)
Andy, this list is from this week (Tuesday, I think). Whatamidoing (WMF) (talk) 17:23, 10 June 2016 (UTC)

The latest list is here: phab:T136674#2394147. Whatamidoing (WMF) (talk) 18:19, 21 June 2016 (UTC)

Yes, I'm still there with Huggle 2. I'll have to keep using it till it chokes . - DVdm (talk) 18:25, 21 June 2016 (UTC)
Why are people stuck on Huggle 2? I think the last commit to it was ~1.5 years ago, and Huggle 3 exists. If Huggle 3 is not usable by some, perhaps we can ask the Huggle authors to do a final point-release of Huggle 2 and fix the HTTPS issues in it? BBlack (WMF) (talk) 14:02, 24 June 2016 (UTC)
@BBlack (WMF): (sorry, a bit late) Why stuck on Huggle 2? Because H3 fails where H2 brilliantly excels: in H3 the horizontal Page and User Editbars don't show the current Page and User status. Well, sometimes they do, but most of the time they don't, so, being unreliable, they cannot be counted upon. Without them, the tool feels unsafe to me. By the way, the user warning levels aren't updated in the icons in the Page history window either:
Difference between editbars in Huggle 2 (top) and Huggle 3 (bottom). Current user warning level is shown in H2, but not in H3. Not shown in User column in Page History window either.
I don't think the Huggle author volunteers have the time or the resources to fix a final point release—see the lack of response at User talk:DVdm#HG. Alas. - DVdm (talk) 08:21, 5 July 2016 (UTC)
FWIW, see Wikipedia:Huggle/Feedback#User status not updated in Huggle 3. With crossed fingers . - DVdm (talk) 09:37, 5 July 2016 (UTC)

Username list updates

Continuing updates to recent lists of usernames still making insecure accesses are being posted every few days here: http://phabricator.wikimedia.org/T136674 --BBlack (WMF) (talk) 14:31, 24 June 2016 (UTC)

Last chance

The API change is scheduled for tomorrow (less than 24 hours from now). If anyone has a bot that might not be fixed, then you might want to run it ASAP, to get one last run before you find out.  ;-)

To recap from above, you can find general information in this e-mail message from May. The list of bots/script users that are known to be affected is listed at phab:T136674. If you need information to figure out which piece is going wrong, then the devs can give you a user agent string and similar details from the logs. Just let me (or BBlack (WMF)) know that you need more details.

Good luck, Whatamidoing (WMF) (talk) 16:22, 11 July 2016 (UTC)

All things considered, we've decided to soften the final cutoff date just for Labs at the last minute. Tomorrow's patch will disable insecure access completely from non-WMF networks, but we'll still allow it from our hosted Labs instances for another full week. The Labs bots that continue using insecure access will be subject to 20% random failure rate (up from 10% this past month) which should make it even more noticeable. If you need help getting a bot's configuration or library updated this week, please reach out. A good place would be the primary phabricator ticket for this issue: http://phabricator.wikimedia.org/T105794 . The patch deploying the above changes tomorrow is: http://gerrit.wikimedia.org/r/#/c/298336/ BBlack (WMF) (talk) 00:55, 12 July 2016 (UTC)
The final patch, to disable insecure access from hosted Labs instances, went out earlier today: http://gerrit.wikimedia.org/r/#/c/299532/ I haven't seen any related problems reported in the last few days, here, VPT, or the related phab tasks, so trust all is well. Let us know if I missed anything. Cheers. Quiddity (WMF) (talk) 05:47, 20 July 2016 (UTC)
Thanks for the update. — xaosflux Talk 11:25, 20 July 2016 (UTC)

Relevant discussion

Template_talk:Emergency-bot-shutoff#ANI_notice might be of interest to editors who stalk this page. Σσς(Sigma) 05:36, 21 July 2016 (UTC)

BattyBot

I have blocked BattyBot (talk · contribs) as an emergency because it seems to be going over large amounts of articles changing {{main}} to {{main article}} for no obvious reason random example. If somebody can explain what's going on, I'll unblock. Ritchie333 (talk) (cont) 19:01, 18 August 2016 (UTC)

@Ritchie333: As stated in the edit summary, the bot did two things:
  1. Changed the incorrect link to Billboard to a piped link to Billboard (magazine) (the primary task)
  2. AWB general fixes, including bypassing template redirects by changing {{main}} to {{main article}}.
Since the bot is working as designed, please unblock the bot. In the future, please just post to User talk:BattyBot, which will stop the bot and make it easier to find your post. Thanks! GoingBatty (talk) 19:09, 18 August 2016 (UTC)
Seems a bit pointless - I mean, "main's" easier to type. And reading the edit summary, "replaced: |work=Billboard → |work=Billboard" looks like it doesn't obviously do anything. Oh well, unblocked. Ritchie333 (talk) (cont) 19:11, 18 August 2016 (UTC)
@Ritchie333: If you would like to discuss your concern about the {{main}} to {{main article}} replacement, please post at Wikipedia talk:AutoWikiBrowser. If you have a suggestion on how I could improve the rest of my edit summary, please post at my talk page. Thanks for unblocking! GoingBatty (talk) 19:36, 18 August 2016 (UTC)

B-bot tagging of FFD candidates

B-bot's first task is to tag non-free images which are orphaned for deletion. When processing these deletion requests I did decline two requests File:Playdead logo.gif and File:Careers360 Logo.png because I had doubts about their copyrightability and listed them at FFD instead. B-bot keeps tagging them though despite a {{bots|deny=B-bot}} tag. I did ask on B's talk page about this but they don't seem to be very active. Is there a way to stop the bot tagging? Jo-Jo Eumerus (talk, contributions) 12:02, 12 July 2016 (UTC)

@Jo-Jo Eumerus: This does appear to be malfunctioning (in that it is not honoring exclusion compliance); I emailed the operator as well. With exclusion compliance broken, our normal response would be to block the bot; right now it appears to only be impacting these 2 pages - I reverted the bot and placed temporary TPROT on those pages to make the bot go away while this is under discussion. If the impact is to only these 2 pages, this should be sufficient for now. Let's give the operator a chance to respond as the rest of the bot edits appear to be productive still. — xaosflux Talk 13:48, 12 July 2016 (UTC)
@B and Xaosflux: Seems like the issue is still there.Jo-Jo Eumerus (talk, contributions) 12:37, 25 July 2016 (UTC)
Given the importance of quickly deleting orphaned non-free images, I don't really see blocking B-Bot as a possibility unless another bot operator adopts the task first. The solution might have to be protecting these pages until their deletion discussions conclude, at least until another solution is in place. ~ Rob13Talk 21:08, 18 August 2016 (UTC)

Your Input on Archive Bot Behavior

Hi there,

What do you think about this edit (and the corresponding edit to the article's talk page). I find access dates by finding the earliest occurrenceof the url in the article. Is that a good idea? What are the standards for bots that add archives? --Tim1357 talk|poke 02:45, 23 September 2016 (UTC)

API Outage

FYI: There is/was an API outage today, see Wikipedia:Village_pump_(technical)#503_errors.2Fpossible_Twinkle_issues. — xaosflux Talk 16:20, 17 October 2016 (UTC)

Conflicting edits between Hazard-Bot and Cyberbot I

As of this edit, It looks like that the two bots have different "opinions" on what a blank template sandbox should look like.

To me, one of the bots should either stop patrolling the template sandbox, or stop doing the task completely if the bot task only patrol that page. NasssaNser (talk/edits) 04:07, 22 October 2016 (UTC)

Pinging operators: Cyberpower678 and Hazard-SJ. — JJMC89(T·C) 05:31, 22 October 2016 (UTC)
Optimally, both bots should be able to get along just fine without arguing about what the content should be, which, as far as I'm aware, was why we have templates (in this case Template:Template sandbox reset) to guide what the content should be. My bot substitutes the aforementioned template, so that if there is a decision to change the content at some point, it could be easily done without my intervention, and especially if multiple bots (as is the case here) perform the same task, both would easily "update" to use the most recent content immediately and simultaneously. Checking the page history, your link is not nearly the first occurrence of the issue, and it had also happened with lowercase sigmabot II and Cyberbot I as well as in this edit). Cyberpower678, would it be okay for you to also adopt usage of that template as well?  Hazard SJ  21:07, 23 October 2016 (UTC)
Sure. The text is predefined values when the script starts.—cyberpowerTrick or Treat:Offline 02:11, 24 October 2016 (UTC)

High false positive rate for User:InternetArchiveBot

As its maintainer's talk page witnesses, User:InternetArchiveBot is currently not working very well. As the recent error reports have not been answered and the bot is still making edits, would it be useful to block it? − Pintoch (talk) 16:36, 22 October 2016 (UTC)

@Cyberpower678: Any comments here? — xaosflux Talk 17:38, 22 October 2016 (UTC)
The bot operator was active and fixing bugs as recently as 14 October 2016. See the talk page history. – Jonesey95 (talk) 19:31, 22 October 2016 (UTC)
I wouldn't call this a high false positive rate. During testing it's been established that the false positive rate is 0.1%. Let's take 0.1% of 40 million links. That's 40,000 links. So it may seem high, but not if you look at the bigger picture.—cyberpowerTrick or Treat:Offline 22:32, 22 October 2016 (UTC)
Ok, good to know. I was interested in this for my own bot. − Pintoch (talk) 07:47, 23 October 2016 (UTC)

C678 how was the 0.1% estimated? -- Magioladitis (talk) 23:02, 24 October 2016 (UTC)

The community tech team and I went through a sample of 1000 URLs that were tested by the bot, during the trial, and found only 1 to be a false positive.—cyberpowerTrick or Treat:Offline 01:59, 25 October 2016 (UTC)

The modification you tried to make was aborted by an extension hook

Good afternoon! When bot edit pages frequently get the error: "The modification you tried to make was aborted by an extension hook". How can it be called? Игорь Темиров (talk) 13:40, 29 October 2016 (UTC)

@Игорь Темиров: this error message may be related to you hitting the abuse filter on a specific project. If this is on a WMF wiki, can you provide the account name, project name, and page you are trying to make the edit to? — xaosflux Talk 14:36, 29 October 2016 (UTC)
@Xaosflux: No, this page. Some similar pages are saved properly, but part with this error. Игорь Темиров (talk) 15:39, 29 October 2016 (UTC)
@Игорь Темиров: I have global abuse-filter access so was able to verify you are not hitting ruwiki's AbuseFilter, however you may be hitting their ru:Special:Log/spamblacklist - you will need to ask a ruwiki administrator to check that for you. — xaosflux Talk 15:55, 29 October 2016 (UTC)


RFCs on citations templates and the flagging free-to-read sources

As bot operators, you might want to follow these discussions

Headbomb {talk / contribs / physics / books} 16:23, 29 October 2016 (UTC)

New adminbot request

There is a new adminbot request open at WP:BRFA. Please see Wikipedia:Bots/Requests for approval/MusikBot II for details. — xaosflux Talk 04:16, 1 November 2016 (UTC)

Problematic edit and confusing minor edit summary by poorly identified bot - GreenC bot / WaybackMedic / Wayback Medic

http://en.wiki.x.io/w/index.php?title=Taser&diff=737265819&oldid=736467529 is tagged GreenC bot and (WaybackMedic 2).

  1. It shouldn't be at all ambiguous whether Wayback Medic or GreenC bot did the edit. Is there a good reason for the discrepancy or should it be fixed?
  2. The edit is problematic for several reasons. Information about archived content that isn't viewable due to IA's robots.txt policy is not useless. It's a) subject to legal discovery at any time, and b) viewable if the robot.txt changes. c) the existence of the IA URL suggests that the content was verifiable d)Deleting it means a normal viewer of the article won't learn that, and e) the source is available at Archive.is : [6]. For these five reasons, the bot should not be removing non-working IA URLs, especially not without checking to see if there's a working Archive.is version.--Elvey(tc) 01:02, 15 November 2016 (UTC)
An observational comment: Archive.is has no API from what I can tell making it nearly impossible for bots to use it.—cyberpowerChat:Offline 01:16, 15 November 2016 (UTC)
Well, then go fix our Archive.is article, User:Cyberpower678; it says they have an API. LOL. --Elvey(tc) 02:04, 15 November 2016 (UTC)
Memento isn't part of Archive.is. I mean that archive.is doesn't have it's own API bots can use, and sucks in general for delivering reliable results. It doesn't preserve HTTP codes making it impossible for bots to determine if it's a good archive or not.—cyberpowerChat:Offline 03:01, 15 November 2016 (UTC)
In my experience it was like 50% soft-404 making it unusable, unless we spam Wikipedia with broken links. Elvey, do you know someone at Archive.is we can speak with about that? -- GreenC 03:26, 15 November 2016 (UTC)
Archive.is seems to be a rather unresponsive one man operation. Without reliable communications, I absolutely refuse to use archive.is beyond the current URL recognition.—cyberpower (unsigned)
I'm confused. I hear you say WM can tell that many archive.is links were broken from the start - and that you need an API to be able to tell if the links work or not. Seems contradictory. What am I missing? --Elvey(tc) 07:52, 15 November 2016 (UTC)
AI uses Memento for their API. The links returned by Memento say the page is available. The page headers at AI show the pages works (status 200). A manual inspection of the page shows 404 (ie. soft-404). Thus the only way to verify a page is working is manually - the API is unreliable, and the page headers are unreliable. -- GreenC 14:22, 15 November 2016 (UTC)
Thanks.--Elvey(tc) 02:04, 15 November 2016 (UTC)
  1. There is no User:WaybackMedic. "WaybackMedic" is the name of the program being run by the bot account User:GreenC bot. The account runs multiple programs.
  2. WM is currently not removing 403 robots.txt links anymore but it is removing other types of 403s, 404s, certain types of 301/302. The reasons are many, some you described, and for other reasons. WM processed over a million links, deleted 30,000 (3%) and I am considering a project to restore 403 robots.txt that are back alive (about 3000 or 0.3%). As for checking other archives, absolutely WaybackMedic does it through Memento's API and it has saved a lot of links that way, mostly WebCite and LOC (see the project's Stats page "New alt archive"). Archive.is a special case and largely unusable by bots, as Cyberpower said. It doesn't prove the link was once verifiable, or even existed - WM found many links that never worked because they were added by bots that didn't have access to the IA API - the links were broken from the start. Furthermore monitoring has shown these non-robots.txt links are permanently dead. Wayback for other reasons (some intentional some due to internal data problems) will delete links unrelated to robots. -- GreenC 03:24, 15 November 2016 (UTC)
  1. Ok. "Wayback Medic is a bot" is what it says at User:Green_Cardamom/WaybackMedic_2 which you link to. Maybe that should change?
  2. Great. Appreciate all the tons of good edits you/your bot does.--Elvey(tc) 07:52, 15 November 2016 (UTC)

Per this discussion and this one on the Village pump (technical), I have raised the possibility of removing Template:Fa top's invocation of the metadata class, which in mobile view has the declaration display:none;, from the div section. This will allow the featured log to be viewed in the "mobile view". However, there was some fear that it might break a bot. The FACBot is okay, and I'm not aware of any other bot that accesses these pages, but it's best to ask first rather than proceeding and seeing what breaks. (Notifying Ian Rose and Bonvol:) Hawkeye7 (talk) 20:14, 1 November 2016 (UTC)

Okay. I have removed it. We'll see if anything breaks. Hawkeye7 (talk) 23:23, 15 November 2016 (UTC)

{{webarchive}} merge

A question has come up at User_talk:GreenC_bot#GreenC_bot if the bot should honor the configuration of the original {{wayback}} template or {{dmy}}. There are good arguments either way, I felt honoring {{wayback}} was the safer course of action as the only purpose of the bot is merger, not creating a new date format change. The bot has now been shut down twice by User:Me-123567-Me who also took it to ANI so I'm starting this thread for discussion. I'm willing to make this change, it's not difficult, but it may have unintended consequences and the bot was not approved for it. -- GreenC 22:21, 15 November 2016 (UTC)

It should honour the original {{wayback}} template, if it has an explicit df parameter. If not, it must honour the date format specified in the {{dmy}} template. Hawkeye7 (talk) 23:20, 15 November 2016 (UTC)
Alright. Sounds reasonable that way. I can do it, including back over the old ones first, but if anyone complains you may be pinged :) ( I'll wait till tomorrow to start for any other comments.) -- GreenC 23:41, 15 November 2016 (UTC)
@Hawkeye7: - this also impacts the merger of {{webcite}} which has a default of dmy when no |dateformat= is specified. If the bot finds no {{dmy}} it will convert to mdy, for better or worse. -- GreenC 00:06, 16 November 2016 (UTC)

Done. 50069 articles originally edited, 10125 had a {{dmy}} and 849 articles fixed (example). -- GreenC 19:29, 16 November 2016 (UTC)

Thank you for making that change. I greatly appreciate it. Me-123567-Me (talk) 21:45, 17 November 2016 (UTC)

The bot maintains both Peer Reviews and Good Article Reassessments. It hasn't been active since November 4 and both PR and GAR hasn't been updated since. The people operating it have said they don't maintain the bot (CBM, Ruhrfisch). Is it possible to get someone to make a new bot or assume control of the bot to fix this situation? GamerPro64 04:33, 24 November 2016 (UTC)

@CBM: and @Ruhrfisch: - if this bot is no longer being controlled it will be considered retired and be soft blocked/de-flagged. If you are still interested in operating this bot please acknowledge. — xaosflux Talk 05:03, 24 November 2016 (UTC)
Thanks for the heads up - I will copy and paste what CBM and I originally posted here:
User:VeblenBot handles many of the routine chores associated with Peer Review. It was developed by User:CBM, and is currently in my care, but neither of us has the time or inclination to run it. Would someone be able to take it over? If so, please reply here. Thanks, Ruhrfisch ><>°° 19:06, 20 November 2016 (UTC)
Much of this is just a task-specific archiving job. It would be possible for someone else to rewrite this in a bot framework of their choice without too much work, instead of taking over the existing code. It's an important task for the Peer Review system, but I can't manage it any longer. — Carl (CBM · talk) 13:05, 21 November 2016 (UTC)
I will also add that VeblenBot does useful work for the Peer Review, Good Article, Featured List, and Featured Article processes, so having someone take it over or duplicate its tasks would be very helpful. Ruhrfisch ><>°° 23:30, 25 November 2016 (UTC)

Wikiget tool

I wrote a tool for myself, found it pretty useful so cleaned it up and posted on GitHub in case anyone might be interested.

It's a unix command-line tool to retrieve lists of article names, such as all pages in a category, backlinks of a template, pages edited by a user during a certain period, etc.. it's generally useful for work with AWB or bots, but probably other things as well in a unix environment. Only dependency is GNU awk and one of wget, curl or lynx.

https://github.com/greencardamom/Wikiget

-- GreenC 01:41, 26 November 2016 (UTC)

Inactive bots with inactive editors to mark as retired

The following bots appear to be inactive, have no planned future tasks, and have inactive operators.

  • Bot, last edit, operator, last edit
  1. User:D6, 20140330, @Docu:, 20140417
  2. User:Wikipedia Signpost, 20060725, @Flcelloguy:, 20070529 (this account is claimed and will be left alone for now)

Baring any objections, I propose to mark these bots as retired and deflag them. They may be reactivated with a future BRFA A longer list will be coming, but wanted to put these initial example out to see if there is any community objection to this approach. — xaosflux Talk 17:14, 25 November 2016 (UTC)

This has been done previously, however inactive bot removal is not specified in the bot policy though I plan to propose an addition to reflect this practice (again - baring any objections here). — xaosflux Talk 17:18, 25 November 2016 (UTC)
  • Strong support, especially in light of the recent compromised account issues. ~ Rob13Talk 07:49, 26 November 2016 (UTC)
  • Strong support, considering the recent account compromises. The older accounts are probably much more vulnerable.—cyberpowerChat:Online 16:35, 26 November 2016 (UTC)
  • Comment re: Signpost account. I am the Signpost editor; I've now blocked that account per the reasons cited. While I am not deeply familiar with the policies around bots, if possible, I'd like to request that the bot flag not be removed; we are currently adjusting our publication process and may want to reactivate the bot soon. (If that would require new approval though, that's OK.) -Pete Forsyth (talk) 19:15, 26 November 2016 (UTC)
    Thank you @Peteforsyth: since you are in control of the account no need to remove the flag right now. The user page should be updated to reflect the current operator (you?) and ancient approval. Please note, the MassMessage extension is generally preferred to manually updating user talk's for "spam"ing people. If it will be used for that same task, no new BRFA is needed (but the MMS flag may be). If it will be used for some other editing task, BRFA is easy enough. — xaosflux Talk 20:05, 26 November 2016 (UTC)
    Thanks Xaosflux, I will update the user page. To be clear, we have indeed been using MasMessage for that purpose (both on enwp and globally) for some time, as documented here. The bot functionality we badly need right now is not for the announcements, but to facilitate the other components of the publication process. -Pete Forsyth (talk) 20:15, 26 November 2016 (UTC)
    If that's not something the bot was previously approved for, a new approval would be needed. ~ Rob13Talk 20:33, 26 November 2016 (UTC)
 Done User:D6 marked retired, deflagged. — xaosflux Talk 01:34, 2 December 2016 (UTC)

Proposed activity requirements for maintaining bot flags

Please see Wikipedia_talk:Bot_policy#Activity_requirements for a proposed amendment to the bot policy. — xaosflux Talk 19:21, 3 December 2016 (UTC)

Inactive bots over 5 years

(See report: Wikipedia:Bot owners' noticeboard/InactiveBots/December2016)

The following bots appear to be inactive, in that they have no edits for 5 years. In some special cases this may be due to a read-only bot that only exists for highapi read access. There are many reasons these could be inactive: tasks may have been moved to other accounts, all one-tasks were one-time and completed, operator may have left wikipedia, etc.

Extended content

Bots that have not edited since 2010 as of 26 November 2016 (UTC). [edit]

Bot Last botedit Operator Last opedit KeepPlease
Wikipedia Signpost 23:10, 25 July 2006 (UTC) Flcelloguy 20070529 20:13, 26 November 2016 (UTC) (see WP:BOWN)
XyBot 13:06, 29 May 2007 (UTC) FF2010 20151201
Alphachimpbot 06:47, 3 June 2007 (UTC) Alphachimp 20160620
Rob110178bot 01:14, 25 September 2007 (UTC) Rob110178 20141101
CloudNineBot 17:17, 3 December 2007 (UTC) CloudNine 20130413
WASDbot 22:10, 20 December 2007 (UTC) WASD@ruwiki 20160831
AMbot 03:58, 8 January 2008 (UTC) After Midnight 20161123 After Midnight 0001 15:42, 26 November 2016 (UTC)
AccReqBot 19:00, 3 March 2008 (UTC) ST47 20161002
HBC AIV helperbot 8 20:35, 13 April 2008 (UTC) Monobi 20080621
Town-bot 20:03, 20 May 2008 (UTC) Hughey 20161121
HBC AIV helperbot4 00:19, 11 August 2008 (UTC) ST47 20161002
AloysiusLiliusBot 03:58, 23 September 2008 (UTC) Ameliorate! 20160423
Nomenclaturebrowser 10:29, 15 October 2008 (UTC) Aaronbrick 20160922
AmeliorationBot 06:28, 8 November 2008 (UTC) Ameliorate! 20160423
CwraschkeDataBot 19:10, 13 November 2008 (UTC) Cwraschke 20081119
Alaibot 02:03, 23 November 2008 (UTC) Alai 20090520
AilurophobiaBot 07:41, 24 November 2008 (UTC) Ameliorate! 20160423
NVS(bot) 18:39, 20 December 2008 (UTC) NonvocalScream 20111203
AWeenieBot 20:08, 13 February 2009 (UTC) AWeenieMan 20081008
UnCatBot 05:26, 28 March 2009 (UTC) nn123645 20131210
MandelBot 23:09, 24 April 2009 (UTC) J JMesserly 20161120 J JMesserly (talk) 16:24, 26 November 2016 (UTC)
SoxBot VII 04:49, 2 May 2009 (UTC) X! 20161125
SoxBot II 09:36, 3 May 2009 (UTC) X! 20161125
SoxBot V 14:00, 15 May 2009 (UTC) X! 20161125
SoxBot VI 18:03, 15 May 2009 (UTC) X! 20161125
OrphanBot 06:43, 14 June 2009 (UTC) Carnildo 20161120
ARSBot 22:00, 12 October 2009 (UTC) ST47 20161002
Werdnabot 09:12, 26 October 2009 (UTC) Werdna 20150511
BAGBot 17:55, 4 January 2010 (UTC) ST47 20161002
ClueBot IV 07:32, 6 January 2010 (UTC) Cobi 20161101
RFC posting script 17:33, 21 February 2010 (UTC) Harej 20161121
HBC NameWatcherBot 03:00, 7 March 2010 (UTC) HighInBC 20161118
AdambroBot 13:06, 29 March 2010 (UTC) Adambro 20121004
Orphaned image deletion bot 00:02, 29 April 2010 (UTC) Chris G 20151217
HBC AIV helperbot2 10:37, 29 April 2010 (UTC) Alphachimp 20160620
SandgemBot 07:07, 13 June 2010 (UTC) Sandgem Addict 20161125
Stwalkerbot 19:08, 13 June 2010 (UTC) Stwalkerster 20161123 [stwalkerster|talk] 18:15, 26 November 2016 (UTC)
Image Screening Bot 03:29, 26 June 2010 (UTC) Johnduhart 20140413
Signpost Book Bot 21:54, 29 June 2010 (UTC) Johnduhart 20140413
PDBbot 17:06, 18 July 2010 (UTC) Emw 20160525
Kakashi Bot 22:40, 11 August 2010 (UTC) AllyUnion 20160811
Commander Keane bot 04:00, 20 August 2010 (UTC) Commander Keane 20161126
BrokenAnchorBot 22:01, 2 September 2010 (UTC) Winston365 20140606
SunCreatorBot 17:20, 5 September 2010 (UTC) Sun Creator 20161124
ShepBot 16:02, 17 September 2010 (UTC) Stepshep 20141121
CleanupListingBot 20:43, 20 September 2010 (UTC) Smallman12q 20130701
KolBot 13:37, 1 October 2010 (UTC) Samuel Wiki 20161125
Muro Bot 21:22, 6 October 2010 (UTC) Muro de Aguas 20150429
PC78-bot 14:24, 8 October 2010 (UTC) PC78 20161031
WikiStatsBOT 05:27, 14 October 2010 (UTC) ThaddeusB 20160407
LawBot 06:26, 17 October 2010 (UTC) MZMcBride 20161125
ClueBot VI 18:05, 20 October 2010 (UTC) Cobi 20161101
Chrisbot 16:33, 17 November 2010 (UTC) Random_user_8384993 20101117
ClueBot 20:10, 2 December 2010 (UTC) Cobi 20161101
OpenlibraryBot 19:21, 3 December 2010 (UTC) Arielbackenroth 20120612
SelectionBot 21:20, 4 December 2010 (UTC) CBM 20161122
WMUKBot 17:30, 7 December 2010 (UTC) Seddon 20161122
718 Bot 04:01, 8 December 2010 (UTC) east718 20161114 east718 | talk | 16:27, 29 November 2016 (UTC)
XLerateBot 00:22, 10 December 2010 (UTC) XLerate 20120328
Smallbot 23:05, 22 December 2010 (UTC) Smallman12q 20130701

I propose marking this bots as retired, and removing their bot flags. All operators will be attempted to be contacted via user talk and can indicate that they want to maintain their bot flag by signing the table. Any bots retired in this manner will be consider de-authorized, but may be reactivated in the future following a successful BRFA. The bot policy does not currently have a provision for activity requirements, but assuming this process has support I will propose a policy update to reflect practice. — xaosflux Talk 14:23, 26 November 2016 (UTC)

As long as we give a warning, I'm fine with this. –xenotalk 14:25, 26 November 2016 (UTC)
100% I'm going to message each operator now - primary goal here is to ensure these accounts are under control still. — xaosflux Talk 14:27, 26 November 2016 (UTC)
Re MandelBot, It has a strong password, and has been inactive for an extended period due to lack of time (many young children now) to spend on Commons- chiefly to normalize categories according to agreed on standards. I expect to continue my work in the future because the work load still is pretty daunting for less technically capable commoners/ unfamiliar with how to create safe custom bot programs. I would strongly prefer that it not be retired- for example if retired I would not be able to promptly respond to the requests from trusted colleagues to assist in reclassifications. I appreciate the concern over misuse and support the idea of retiring such accounts especially if they could be compromised. J JMesserly (talk) 16:18, 26 November 2016 (UTC)
No worries, we're making sure everyone still wants them - marked for keep. — xaosflux Talk 16:52, 26 November 2016 (UTC)
I'm okay if User:SunCreatorBot is deactivated. Regards, Sun Creator(talk) 17:46, 26 November 2016 (UTC)
HBC NameWatcherBot is defunct and has been replaced with another bot. It no longer needs its flag. HighInBC Need help? {{ping|HighInBC}} 21:06, 26 November 2016 (UTC)

We 've done this before. So.. proceed! -- Magioladitis (talk) 22:56, 26 November 2016 (UTC)

I'm listed for LawBot. Flag removal is fine with me. Two down, three to go! --MZMcBride (talk) 14:39, 2 December 2016 (UTC)

Request for BAG membership (bot approver)

Hello! I have offered to help with the WP:BRFA backlog as a bot approver. This procedural notification is to make the community aware that a formal request is open for your consideration. Your input is welcomed at Wikipedia talk:Bot Approvals Group#BAG Nomination: MusikAnimal. Regards MusikAnimal talk 00:55, 9 December 2016 (UTC)

Bots that never ran to deactivate

Here is another batch of bots, these ones have never made an edit so were missed in the prior report. All have operators have not edited in 5+ years. Operator talk messages being left. Barring any objections, will mark as retired and deflag in a week:

xaosflux Talk 18:00, 3 December 2016 (UTC)

Thanks. I have no quibbles with AlekseyBot or JatBot, since they were at least intended to edit. The other two weren't -- so is there not a very small chance that they are continuing to perform some invisible yet helpful function? - Jarry1250 [Vacation needed] 18:59, 3 December 2016 (UTC)
If their operators speak up I'm fine with keeping - and if they dont but one day come back getting access again shouldn't be a hurdle. — xaosflux Talk 16:04, 6 December 2016 (UTC)

Which bot?

I know there was (at least) one bot for searching copyvios. Does anyone know which is that bot? XXN, 11:48, 14 December 2016 (UTC)

XXN at least one is Wikipedia:Bots/Requests for approval/CorenSearchBotxaosflux Talk 12:25, 14 December 2016 (UTC)
Thank you, Xaosflux. I remembered another things with this function, but they are not really bots, rather than tools: [7], [8]. --XXN, 12:59, 14 December 2016 (UTC)

Inter-wiki bot approval

Hi, is there a mechanism or bot approval for work across all language wikis? I recently answered a request for some low-volume high-accuracy non-controversial changing URLs for example sound.westhost.com to sound.whsites.net .. these URLs exist on other language wikis and should be fixed, but I don't have bot approval there (I don't think?). For example DE has about 8 and FR has about 4 etc.. it's too much work to go through each language manually, and not worth the effort to apply for bot approval in every language that needs it for this one task. -- GreenC 15:50, 14 December 2016 (UTC)

@Green Cardamom: not as far as I can tell - but please chime in at meta:Talk:Bot_policy#Reconsidering_the_policy for a current discussion about non-controversial expansions for the global bot policy. Note: most of the "large" wiki's opt-out of that policy. For something like '4 edits on frwiki' - I'd just use my editor account. — xaosflux Talk 16:09, 14 December 2016 (UTC)
Ok I'll check out the discussion. There are many languages so it might be a lot of edits in total, this is just an example case. We need an International body who can approve some low impact non-controversial bots, like URL changes. -- GreenC 19:43, 14 December 2016 (UTC)

Request for Bot Approval Group membership seeking input

Hi everyone. I am currently requesting to join the Bot Approval Group, and notification on this page is required. Feel free to comment here if you would like to ask questions or discuss the request. Wugapodes [thɔk] [ˈkan.ˌʧɻɪbz] 23:57, 18 December 2016 (UTC)

BAG reconfirmation

A bot approvals group member reconfirmation discussion is now open at Wikipedia:Bot Approvals Group/nominations/Magioladitis 2. Please feel free to review and comment. Thank you, — xaosflux Talk 13:25, 20 December 2016 (UTC)

A review of Yobot's operation

Per the bot policy, a request to review Yobot's tasks for continuing authorization has been brought up at: Wikipedia_talk:Bots/Requests_for_approval#Request_to_modify_Yobot_authorization. Please review and comment if you are interested. Thank you, — xaosflux Talk 13:33, 20 December 2016 (UTC)

Removal of data, and Wikidata

Please see Wikipedia:Village pump (policy)#Removal of data, and Wikidata, which suggests a change to bot policy. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:29, 12 January 2017 (UTC)

General question about semi-automated editing

A couple of years ago, I learned to use Macro Express at work (it's a simple keyboard-and-mouse macro), and having need for the same functionality, I bought it a few days ago. Even when I was using it frequently, it often made seemingly random mistakes, so even in my own offline work, I tend to instruct it to wait for an "okay" from me before it saves anything.

I've thought of using it when doing some editing here (e.g. if I've uploaded a bunch of photos that should be added to articles, I set up a macro to copy the files onto the relevant articles), using the following process: the script loads a page, performs the specified modifications to the code, prompts me to approve or reject the edit, and (if I approve it) hits "Save changes" and goes to the next page, or (if I reject it) the macro stops until I reactivate it. I'd like some input:

  1. Because I'd require the macro to wait for my approval on each edit before saving, do I understand rightly that this would be considered semi-automated, comparable to AWB?
  2. It seems that AWB requires no permission to use on specific tasks: you just get added to the approved-users list, and there's no requirement that you get anything comparable to a WP:BOTR before using it, as long as the task won't be controversial. Would that be true of my situation?
  3. Would this kind of process require me to register a separate account, or would it be okay to run on my main account?
  4. Would it be necessary to use an edit summary that mentions that it's not a fully manual edit?
  5. Do we know of anyone in the past who's used keyboard-and-mouse macros for automated or semi-automated editing, and/or anyone who's currently using them? It would help if I could read about their experiences, or ask them if they're still active on-wiki.

I don't think you'd have a way of noticing anything unusual with what I'm thinking of doing, unless you checked my contributions and observed that the edits are going too fast to be manual; that's the only reason, aside from a desire to abide by relevant policy, that I'm bringing this up. Thanks for your input! Nyttend (talk) 00:46, 2 January 2017 (UTC)

@Nyttend: In order:
  1. Yes. If you're approving every single edit, it is considered semi-automated. You still must edit at a reasonable rate. See WP:BOTASSIST.
  2. Yes, assuming reasonable edit rates and no mass changes. If you're going to make identical changes on a massive scale which would require little oversight, you may wish to consider WP:BRFA and the use of a separate bot account with a bot flag.
  3. Again assuming reasonable edit rates and no changes on a massive scale, you could edit from your main account. You may also set up a legitimate alternative account if you prefer, which may be helpful to segregate your edits using this script for review. That's up to you. Note that whatever account you edit from, you are responsible for every edit made using the semi-automated script.
  4. It would be strongly encouraged to aid in transparency. As per WP:BOTASSIST, "Contributors using assisted editing tools may wish to indicate this, if it is not already clear, in edit summaries and/or on the user page or user discussion page of the account making the contributions."
  5. I don't, but others may be aware of this. It isn't normal, but someone must have tried it in the past.
Basically, the goal is to not clog up recent changes, be transparent in how your script operates, and take full responsibility for every edit you make. The rest tends to follow from common sense, but please do ask if you have further questions. ~ Rob13Talk 00:58, 2 January 2017 (UTC)
Thanks for the input (of course I'd welcome additional input from anyone else, too); what you said is what I was guessing, but I've never really paid attention to automation-related things (aside from making bot requests), and aside from the nominate-for-deletion links in the Commons toolbar and the update-this-page link at WP:NRHPPROGRESS, I don't think I've ever made a scripted edit. In my photos example, the process would go as follows:
  1. Macro starts with a Windows Notepad list of filenames and URLs
  2. Take a URL (http://en.wiki.x.io/w/index.php?title=Pagename&action=edit, not http://en.wiki.x.io/wiki/Pagename), activate Internet Explorer, put the URL into the browser bar, go to the page, and wait a few seconds for the page to load
  3. CTRL+F to find the spot in the infobox where the photo goes
  4. Back to Notepad, copy filename, back to IE, paste filename
  5. Six tab characters, hit "enter" for "Show changes", wait a few seconds for the page to load, prompt me to decide whether the change is good, and do nothing until I approve or reject the edit
  6. If I approve it, paste an edit summary (something like "Adding photo; semiautomated edit using Macro Express"), four tab characters, hit "Save changes"
  7. Wait a few seconds for the page to load, back to Notepad, get the next URL, and start over.
Does this sound likely to be too fast? I do this kind of task manually anyway on occasion, and always in small batches; go to [9] and look for Burleigh (Concord, North Carolina), together with a few edits immediately before it, for a typical example. Nyttend (talk) 01:11, 2 January 2017 (UTC)

It looks fine to me. -- Magioladitis (talk) 01:42, 2 January 2017 (UTC)

@Nyttend: There's looking fine and actually acceptable. Even though it's already covered by consensus, I would strongly suggest securing a BRFA for this task. I would also suggest that if you're doing this from some proceduraly generated list, why not format it to be done with AWB or get a Bot Author to craft the script for you to drive on (i.e. why are you re-inventing the wheel). I would also suggest making sure you are crystal clear where you're getting data from and where you're inserting this data so that other editors have less landmass to make objections on. Hasteur (talk) 01:48, 2 January 2017 (UTC)
What data? I'm talking about uploading a bunch of images to Commons myself and then using the macro to add these files to specific articles; I'd manually generate the list and just use the macro to do the actual adding. Nyttend (talk) 02:02, 2 January 2017 (UTC)
PS, on the "why" question — I don't want to bother someone else whenever I want to save a little time on inserting a dozen or twenty images into articles (I upload a lot; Commons:Category:Files by User:Nyttend has nearly twenty thousand images), I can save the macro so I won't have to write it anew the next time I do this, I use this macro software at work (and further use makes me better and better at it) as well as potentially on-wiki, so I can use it easily, while I have no idea how AWB works and absolutely no reason to use it on- or off-wiki, so learning to use it for this kind of task would be a waste of time. Nyttend (talk) 04:17, 2 January 2017 (UTC)

Anyone tempted to use this kind of software might like to look at AutoHotkey, a similar, but free, utility for Windows. I have blogged about one aspect of its use, but it's far more powerful than that suggests. I've also just started Wikipedia:AutoHotkey to faciliate discussion. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:46, 12 January 2017 (UTC)

I use AutoHotkey all the time, but merely to enter strings like "{{subst:uw-|}} - ~~~~" and "~~~~" (having no "~" on my non-US keyboard), and, normally needing the right Alt-key to type "[", "]", "{", "}", I have redefined less used keys for them. Very nice and free tool, and capable of much more, no doubt. - DVdm (talk) 14:32, 12 January 2017 (UTC)

WP:BOTPOL update

I've updated the WP:BOTISSUE section of our bot policy. I don't believe the changes to be controversial, but discussion of the changes is certainly encouraged at Wikipedia_talk:Bot_policy#Changes_to_.27dealing_with_issues.27_section. Headbomb {talk / contribs / physics / books} 16:18, 20 January 2017 (UTC)

Question about bot requests

Before posting at WP:BOTREQ, is there an easy way to see if a bot already does a task?

I have in mind a bot that would tidy up something specific in mainspace (spot duplicate use of citations and tidy them with a refname labelled on the first use) and I don't know how to check if one exists already and I'm not sure it's a great idea to experiment in mainspace to test if it'd get fixed!

Help gratefully accepted.

--Dweller (talk) Become old fashioned! 10:18, 26 January 2017 (UTC)

@Dweller: We don't have a central list of tasks, so you'd have to search BRFAs and/or bot userpages. AWB does this in some cases as part of its general fixes, see WP:GENFIXES#Duplicate Unnamed References (DuplicateUnnamedReferences). I don't think there is a bot that is approved to do this. — JJMC89(T·C) 19:00, 26 January 2017 (UTC)
OK, thanks. I'll post a request then, because it seems a useful task. --Dweller (talk) Become old fashioned! 19:09, 26 January 2017 (UTC)

Cydebot running unapproved task

Cydebot is removing links to deleted categories from userpages and archives despite the fact that it does not have permission to do that. Cyde is rarely onwiki. talkpage contribs (((The Quixotic Potato))) (talk) 03:01, 28 January 2017 (UTC)

@The Quixotic Potato: Wikipedia:Bots/Requests_for_approval/Cydebot_4. ~ Rob13Talk 03:04, 28 January 2017 (UTC)
@BU Rob13: I have read that, and it does not give him permission to do this. (((The Quixotic Potato))) (talk) 03:14, 28 January 2017 (UTC)
@The Quixotic Potato: Every category he removed was a result of a category listed at WP:CFD/W, which is precisely what the approval allows. There are no exceptions in the approval for any particular namespaces. ~ Rob13Talk 03:16, 28 January 2017 (UTC)
@BU Rob13: No, the page you linked to does not allow him to remove links to deleted categories from userpages and archives in userspace (or elsewhere). He is allowed to remove categories though. (((The Quixotic Potato))) (talk) 03:18, 28 January 2017 (UTC)
How so? His bot is approved to handle category deletions listed at WP:CFD/W. This involves orphaning. The bot has run for about a decade doing this. You'll need to point to exactly where this exception for userpages is stated. (edit conflict) "He's allowed to remove categories though" ... that's what those diffs show the bot doing? ~ Rob13Talk 03:20, 28 January 2017 (UTC)
@BU Rob13: Burden of proof is on you. I am not saying there is an exception for userpages. "Automatic implementation of category deletions" is allowed. Removing references to deleted categories is outside of that scope. The diffs show that the bot has been removing wikimarkup that puts a page in a category, which is different from deleting the category itself. We have, for example, a Category:Wikipedia_bots. Any admin can delete that category, but deleting that category is not the same as editing every page that is listed in that category and removing the wikimarkup that places that page in that category. Another example, this time of a category that has been deleted: Category:Wikipedian_sex_workers. (((The Quixotic Potato))) (talk) 03:23, 28 January 2017 (UTC)
If you're going to suddenly accuse a bot of operating without approval after it's been running this task for what...10 years now, you better come with some pretty solid proof instead of telling other people the burden is on them. Anyways, I dug up the approval, see User_talk:Cyde/Archive014#Deletion_summaries, fun stuff. Legoktm (talk) 03:44, 28 January 2017 (UTC)
@Legoktm: That is before the bot started doing this stuff AFAIK, and the task that was approved in that section is a different task. (((The Quixotic Potato))) (talk) 03:53, 28 January 2017 (UTC)
Update: Its even worse than I imagined, it has been doing this stuff without community consensus for over 10 years now [10]. I'll probably have to start an RfC or something like that to stop it from editing user(talk)pages and archives in userspace. Sigh. That is a huge amount of pointless edits (even more pointless than cosmetic edits which are explicitly disallowed)! (((The Quixotic Potato))) (talk) 03:59, 28 January 2017 (UTC)
You're free to start an RFC if you really want to, but it appears from the link Legoktm posted that the bot really was approved to perform edits like the one you link. As for the edits being "pointless", I note the pages being left in red-linked categories would clutter Special:WantedCategories. Anomie 14:45, 28 January 2017 (UTC)
Special:WantedCategories is AFAIK one of the least useful special pages in MediaWiki. When you look at the pageview stats and the associated talk page Wikipedia_talk:Special:WantedCategories it becomes clear that almost no-one is using it. If a MediaWiki dev wants to make it more useful then it would probably a good idea to exclude the userspace, or better yet, allow people to filter by namespace. The BAG has approved the bot for a different task, and members said they do not have a problem with the bot (it had been running an unapproved task for a while by then), but there is no community consensus that has approved the task I am complaining about. (((The Quixotic Potato))) (talk) 14:53, 28 January 2017 (UTC)
Deletion of categories implies removing them from use, just like deletion templates also means removing them from use. Feel free to review Wikipedia:Bot_policy#Dealing_with_issues, but I suspect that it will conclude with people not seeing the issue / confirming the bot's long-established and supported behaviour. Headbomb {talk / contribs / physics / books} 15:18, 28 January 2017 (UTC)
One day he just started doing a bunch of stuff without prior approval, and the handful of BAG members don't have a problem with the bot, but that is not the same thing as having this specific task approved... The community hasn't approved this particular task. Over the years it has annoyed many Wikipedians by editing their userspace without permission or consensus. Deleting a category is not the same as editing every page that is listed in that category and removing the wikimarkup that places that page in that category, and one does not imply the other. (((The Quixotic Potato))) (talk) 15:21, 28 January 2017 (UTC)
Again, see Wikipedia:Bot_policy#Dealing_with_issues. Headbomb {talk / contribs / physics / books} 15:26, 28 January 2017 (UTC)
It says: "you should contact the bot operator directly via their user talk page". I have. (((The Quixotic Potato))) (talk) 15:28, 28 January 2017 (UTC)

For what it's worth, Wikipedia:Categories for discussion/Administrator instructions currently says (step 6.6) Go to the old category and check "what links here". If there are links to it within templates, articles, redirects or other category pages, then edit these to link to the new name instead, or remove the links, as appropriate. ... Also update manually-edited pages such as "Categories" sub-pages of portals and active WikiProjects, but ignore "article alert" subpages, user pages and talk pages. (emphasis added).

In 2008, that same page said (step 7.3) All references to the category in articles (and other pages) must be changed to the new name, or removed, as the case may be. So in 2008, when Cydebot was approved for "Automatic implementation of category deletions as a result of listings on Wikipedia:Categories for discussion/Working", it was following the 2008 instructions.

The latter version was modified in 2011 to add "Article alert" subpages, user pages and talk pages can generally be ignored. I was unable to find a corresponding discussion in the talk page archive for that page, but it may have happened somewhere.

It seems possible that the bot was approved following the old instructions, and then the instructions changed. It is unclear whether there was consensus for that change in the instructions. – Jonesey95 (talk) 15:52, 28 January 2017 (UTC)

Indeed, removing categories from other people's userspace (redlinked or not) is discouraged because it annoys people. Heck, editing other people's userspace is discouraged unless you have a good reason to do so. Of course most people are happy if you fix a typo or something like that, we always have to use common sense, but deleting redlinked categories is a controversial task that cannot be done without explicit community approval. So far we haven't found any evidence that the bot was approved by the community for the task I am objecting to (only that a handful of BAG members don't mind the fact that it is doing other tasks), but it is theoretically possible that it was at some point in space and time. It seems unlikely, and it doesn't really matter. If Cyde is willing to stop this particular task then we won't have this problem in the future, and if Cyde is not or does not respond then I'll probably have to start an RfC to stop the bot from editing user(talk)pages and archives etc. in userspace. (((The Quixotic Potato))) (talk) 16:11, 28 January 2017 (UTC)
Jonesey95@, seems to me this is am instruction to leave links to deleted categories alone (e.g. don't touch [[:Category:No longer exists]]), not an injunction against the removal of a deleted category from AALERT pages/user pages (e.g. [[Category:No longer exists]]). Either way, if this is where the dispute comes from, then I suggest starting a discussion at Wikipedia:Categories_for_discussion/Administrator_instructions to clarify what the desired behaviour is. Headbomb {talk / contribs / physics / books} 16:28, 28 January 2017 (UTC)
I agree, 6.6 is clearly about removing links to the category, e.g. [[:Category:Deleted category]]. Step 6.1 seems to be the replacement for the 2008 instructions, where it says list it under the appropriate section of Wikipedia:Categories for discussion/Working which triggers this very bot to do its thing. Anomie 17:10, 28 January 2017 (UTC)

So after reading this lengthy debate, my curiosity was roused to see what all the fuss was about. Wikipedia:Categories for discussion/Log/2016 December 23 seems to explain it. A whole bunch of "Wikipedian who..." categories were deleted because they are in poor taste and inappropriate for an encyclopedia. Wikipedian Jedi Knights that poop with a demented sense of humor, etc. – I'm glad they're gone. What would the point be to just turn the blue links to red? These were all deemed to violate WP:USERCAT. Potato, good bot operators are a rare and precious commodity, and shouldn't be bothered with such concerns. wbm1058 (talk) 03:56, 1 February 2017 (UTC)

Proposal in the Magiodilitis arbcom case

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.


I've made a proposal on the ARBCOM workshop page concerning how to deal with the Magiodilitis situation. Input by WP:BAG, bot owners, and the community at large is welcomed. Even if the proposal doesn't pass, some other ideas can be of interest, especially to WP:AWB/WP:CHECKWIKI people. Headbomb {talk / contribs / physics / books} 13:22, 3 February 2017 (UTC)

#1 may be a bit too strict since Magioladitis uses AWB for a lot of things. If something like Yobot 28 that doesn't seem to do anything related to the current controversy gets (re?)approved, IMO that should probably be ok. Anomie 13:52, 3 February 2017 (UTC)
Right you are. I've tweaked it for a restriction against genfixes/checkwiki fixes rather than any bot fixes. Headbomb {talk / contribs / physics / books} 14:03, 3 February 2017 (UTC)
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Wikipedia:Exclusion compliant

Wikipedia:Exclusion compliant, currently a redirect to Template:Bots, has been nominated for deletion at Wikipedia:Redirects for discussion/Log/2017 February 5#Wikipedia:Exclusion compliant. You are invited to comment at the linked discussion. Thryduulf (talk) 00:36, 6 February 2017 (UTC)

Related to this, I've made a request at WT:BOTPOL asking for someone to write a short explanation of the exclusion-compliant concept. Your assistance would be quite helpful. Nyttend (talk) 04:32, 6 February 2017 (UTC)

Voting time for devs

The new mw:Developer Wishlist is in the voting phase. I believe that all bot coders are considered 'devs' for the purpose of voting rights, so many of the regulars on this page qualify.

This focus of this wishlist is on things that make it easier to be a MediaWiki developer – stuff for coding/reviewing/developing, not stuff for editing Wikipedia. If there are any proposals there that would make your bot work easier, then please vote for them. Whatamidoing (WMF) (talk) 18:56, 8 February 2017 (UTC)

Ops to break wikis, maybe in April

While I'm here anyway: remember m:Tech/Server switch 2016? It's probably going to happen in April, with 14+ days between the two editing interruptions. This is super-early notice, so it may be delayed (it will not happen any sooner), but please make a note that there's a chance that bots will need to be re-started. The official schedule will probably be on the same page as last year, at wikitech:Switch Datacenter. Whatamidoing (WMF) (talk) 21:14, 8 February 2017 (UTC)

Misconversion of diacritics by Legobot

Please see User talk:Legobot#Misconversion of diacritics. --Redrose64 🌹 (talk) 12:35, 16 February 2017 (UTC)

A week or two ago, we wrote this little handy guide for WP:BAG members. It mostly details best practices, and act as a general resources for BAG members. BAG members had a chance to give its opinion on it, but bot owners (and the community at large) are certainly welcomed to suggest improvements to the guide, or point out some things that might be relevant to BAG members that we don't already mention. The discussion can be joined at WT:BAGG.

Thanks for any feedback you might have. Headbomb {talk / contribs / physics / books} 03:13, 25 February 2017 (UTC)

Hi, for quite a while now I haven't had time to properly maintain a few scripts that perform tasks related to RfCs and other ones related to GAN. These PHP scripts are badly in need of love (plenty of unresolved bug reports on Legobot's talk page) and someone who has time to actually debug when something goes wrong:

I'm currently running this on Tool Labs, so I can help with moving the databases over to a new shared account or whatever you'd prefer. Thanks, Legoktm (talk) 20:59, 10 February 2017 (UTC)

A new bot maintainer would be very welcome for the GAN side of things. We've had false failure notices for ages for some of the nominations that pass, and would really like to update the list of GA subtopics so that it matches the latest changes at GA, but we can't without a bot maintainer who has time to deal with requests. (There aren't many requests, but minor things do come up from time to time.) A huge thank you to anyone who takes this on. BlueMoonset (talk) 20:38, 20 February 2017 (UTC)
@Legoktm: I couldn't catch you on IRC (yet), but samtar agreed with me to take over the scripts. How would we go now? Dat GuyTalkContribs 18:01, 25 February 2017 (UTC)

WP:AN content

Moved from WP:AN

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.


User:JJMC89 bot is running through articles that use infobox settlement coordinates parameters and switching to coord. I'm fine with the purpose of the bot, but a significant percentage of the edits made by the bot have been removing heading lines / comments in the infoboxes, such as in this edit. With several dozen to more than 100 parameters used in any given article, these headings in infobox settlement serve a rather useful function and are being removed simply because of a known bug that User:JJMC89 acknowledges, but refuses to address. Repeated requests to fix the problem or shut down the bot left at the user's talk page have been brushed off or simply ignored. WP:Bots suggests WP:AN as the forum for dealing with bots that are not following the "harmless" standard specified by WP:Bot policy.
Any help in resolving this matter will be appreciated, both to stop / fix JJMC89 bot and to have User:JJMC89 fix the articles he has damaged. Alansohn (talk) 02:49, 10 February 2017 (UTC)
User:JJMC89 has responded after all, with the clear indication that he refuses to correct the knowingly defective bot: "As far as I am concerned, there is no problem. Therefore, there is nothing to fix." Alansohn (talk) 02:54, 10 February 2017 (UTC)
  • Again, I have no objection to the bot, but unless the removal of headings / comments is intentional, the bot is malfunctioning. The premise that knowingly removing material from articles is acceptable simply because it doesn't change how the page renders is a blatant misrepresentation of policy. The bug can be fixed and should be fixed and the bot should not be re-enabled until the problem is addressed. Alansohn (talk) 04:34, 10 February 2017 (UTC)
@Beeblebrox: Are you OK with the block being lifted on condition that Task 7 (see above) is suspended until this issue is resolved? — xaosflux Talk 04:40, 10 February 2017 (UTC)
@Alansohn: Beeblebrox may be offline, are you ok with this? — xaosflux Talk 04:55, 10 February 2017 (UTC)
If I read correctly, that makes sense, but I acknowledge my ignorance of the minutia of bots. Alansohn (talk) 05:21, 10 February 2017 (UTC)
OK, Unblocking, but this task (that deals with the coordinates in infoboxes is disabled - needs more disuscsion) - will venue change this to BON for follow up. — xaosflux Talk 05:25, 10 February 2017 (UTC)
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Issue 1: editing HTML comments

  • @JJMC89: to get to the core of this complaint - it does not appear that removing non-coordinates related HTML comments is part of the task approval as referenced in this edit. If you were making this edit by hand - would you have had a coordinates related reason to touch that comment? — xaosflux Talk 05:38, 10 February 2017 (UTC)
    @JJMC89: Can you provide any input to this discussion? — xaosflux Talk 19:29, 10 February 2017 (UTC)
    The HTML comment in the example edit is part of a coordinates parameter value (|longEW= W\n\n<!-- Area/postal codes & others -->\n). The removal of coordinates parameters is approved; therefore, the bot was not malfunctioning, and the edit is within the scope of the approval. Whether a human editor would remove the HTML comment (obviously not required to be removed) has no bearing on whether or not the bot was malfunctioning or acting outside of its approval. — JJMC89(T·C) 03:02, 11 February 2017 (UTC)
    @JJMC89: not sure what I'm missing - in the example edit, <!-- Area/postal codes & others --> appears to be an instructional heading for the next section - not a footer of the "General information" section. Why do you think it is part of that section? — xaosflux Talk 03:45, 11 February 2017 (UTC)
    That is what editors intend it to be. However, as far as the template is concerned it is part of |longEW=. To quote Anomie, [it] is part of the value being passed to the removed coordinates parameter, it's not an independent thing that the bot is additionally removing and the bot isn't malfunctioning by removing it. I never said I thought it was part of any section. — JJMC89(T·C) 05:00, 11 February 2017 (UTC)
    The question here is not is the comment technically part of the parameter, thus technically authorized by the BRFA, thus not a bug? but rather is removal of the comment desired, meant to be authorized by the BRFA, and the bot behaviour within consensus?. The answer to the former matters little. And the answer to the latter is a clear no. So fix the issue and resume editing once that's done. It could be as simple as having some regex to find <!-- Comment -->, replace with |TEMPJJMC89BOT=<!-- Comment-->, apply coordinate logic, then find |TEMPJJMC89BOT=<!-- Comment--> replace with <!-- Comment -->. Or you could do skip conditions when it finds comments and semi-automate those cases. Headbomb {talk / contribs / physics / books} 10:23, 11 February 2017 (UTC)
    My comments have only been to address the former since it was claimed that the bot has a known bug and was malfunctioning, neither of which are accurate. Addressing the latter, changing the bot's code to retain HTML comments, is something that I have not commented on here previously. IMO, retaining HTML comments is not desired. Retaining HTML comments would result in this example edit:
    {{Example
    ...
    |latd         = 1<!-- Latitude degrees -->
    |latm         = 2<!-- Latitude minutes -->
    |lats         = 2<!-- Latitude seconds -->
    |longd        = 4<!-- Longitude degrees -->
    |longm        = 5<!-- Longitude minutes -->
    |longs        = 6<!-- Longitude seconds -->
    
    <!-- "header" comment -->
    ...
    }}
    
    to
    {{Example
    ...
    |coordinates  = {{coord|1|2|3|N|4|5|6|E}}
    <!-- Latitude degrees --><!-- Latitude minutes --><!-- Latitude seconds --><!-- Longitude degrees --><!-- Longitude minutes --><!-- Longitude seconds --><!-- "header" comment -->
    ...
    }}
    
    — JJMC89(T·C) 19:53, 11 February 2017 (UTC)
So how about not touching lines that don't begin with one of the parameters you are converting? — xaosflux Talk 20:28, 11 February 2017 (UTC)
The bot does not operate on lines of text since infoboxes (or parameters) are not guaranteed to be on multiple lines. (The removal is tpl.remove(param).) (\n(?:\s*<!--.*?-->)+) can be used to capture and reinsert giving
{{Example
...
|latd         = 1<!-- Latitude degrees -->
|latm         = 2<!-- Latitude minutes -->
|lats         = 2<!-- Latitude seconds -->
|longd        = 4<!-- Longitude degrees -->
|longm        = 5<!-- Longitude minutes -->
|longs        = 6<!-- Longitude seconds -->

<!-- "header" comment -->
...
}}
to
{{Example
...
|coordinates  = {{coord|1|2|3|N|4|5|6|E}}

<!-- "header" comment -->
...
}}
— JJMC89(T·C) 19:04, 12 February 2017 (UTC)
  • I too do not see the removal of such comments as authorized for this or any other bot, or even remotely likely to garner consensus both in general and in this specific instance. It is not because something doesn't change the appearance of the page that this something can be done. Task 7 can resume iff the bot leaves such comments alone, or consensus has been demonstrated for their removal. Headbomb {talk / contribs / physics / books} 11:45, 10 February 2017 (UTC)
  • Syntactically the comment being removed in the one provided diff is part of the value being passed to the removed coordinates parameter, it's not an independent thing that the bot is additionally removing and the bot isn't malfunctioning by removing it. However, if people find these comments useful it would be a good thing for the bot to recognize them and re-insert them in the semantically appropriate location if this is a common-enough situation to be worth caring about (what is the actual "significant percentage" alluded to?).

    I also don't see any major communication issue here. Perhaps the botop could have been less dismissive of the complaint, but it appears that until it was brought to AN it was just one editor hyperbolically complaining, and the bot op didn't resist the determination by uninvolved editors that the task should be stopped. Anomie 14:14, 10 February 2017 (UTC)

    1/100+ templates that the bot has started/completed processing have a HTML comment that could be possibly worth retaining. — JJMC89(T·C) 03:02, 11 February 2017 (UTC)

What is needed for this task to be reactivated? There are over 300,000 pages that need attention from this bot task (295,000 now, plus more templates that have not had tracking added yet). – Jonesey95 (talk) 00:17, 22 February 2017 (UTC)

@Jonesey95: @JJMC89: assuming you are not willing to stop removing content that is unnecessary for the coordinates conversion than someone else will have to close this to override the complaints if you want to continue. If you will will make changes to avoid this then you have addressed the complaint and this should be able to be resumed. What do you want to do? — xaosflux Talk 22:41, 1 March 2017 (UTC)
I am not the bot's operator. – Jonesey95 (talk) 23:06, 1 March 2017 (UTC). [edited to add: Ping JJMC89Jonesey95 (talk) 00:16, 2 March 2017 (UTC)]
Sorry, wrong ping - fixed. — xaosflux Talk 23:25, 1 March 2017 (UTC)
I provided two possible solutions weeks ago. Either could be implemented. — JJMC89(T·C) 04:20, 2 March 2017 (UTC)
@JJMC89: the second option appears to resolve the original complaint. I've remove protection from User:JJMC89 bot/shutoff/InfoboxCoordinatesParametersMigrator - after you make adjustment and test go ahead and reactivate. — xaosflux Talk 04:43, 2 March 2017 (UTC)

Issue 2: undo's

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.


  • @Alansohn: while your first issue (issue 1) above is being discussed there is also your note about undoing certain edits - I suspect that the scope of this will be very large and undo's just to restore html comment code may not be useful -- thoughts? — xaosflux Talk 05:42, 10 February 2017 (UTC)
    Regardless of the closure on Issue 1, I don't think this is warranted and could result in more trouble. — xaosflux Talk 22:42, 1 March 2017 (UTC)
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

RCStream will be removed on 7 July 2017

I don't know if any of you depend upon this service. There's more information from User:Ottomata here: http://lists.wikimedia.org/pipermail/wikitech-ambassadors/2017-February/001557.html Whatamidoing (WMF) (talk) 23:02, 6 March 2017 (UTC)

Editing outages on Wednesday, 19 April 2017 and Wednesday, 3 May 2017

wikitech:Server switch has new dates: Tech Ops is planning for editing outages on Wednesday, 19 April 2017 and Wednesday, 3 May 2017 (two weeks later). The team has not settled on the exact time yet, but it will probably be between 14:00 and 16:00 UTC (afternoon in Europe/Africa and morning in the Americas).

This is a repeat of the work done last April (only, they hope, a little faster and more reliably). Some bots may need to be restarted after the wikis come back online, so please mark your calendars. Whatamidoing (WMF) (talk) 20:31, 13 March 2017 (UTC)