Talk:Title blacklist/Archives/2020

From Meta, a Wikimedia project coordination wiki
Jump to navigation Jump to search

José Rafael Cordero Sánchez

Hi Billinghurst. pt:José Rafael Cordero slipt through the filter, could you tweak it again, please? Thanks, Sam Sailor 11:29, 1 February 2020 (UTC)

@Sam Sailor: Yes check.svg Done addition 11:34, 1 February 2020 (UTC)
This section was archived on a request by:  — billinghurst sDrewth 07:25, 10 February 2020 (UTC)

Suggested regex

Please add the following due to persistent LTA at enwiki

.*hen.*(?:at|otline).*lin.* <newaccountonly>

The last fifty users this regex scanned on gUser returned two false positives (Zhenya vatulin and Peter.thenatberlin.de), so this might require some significant tweaking. ミラP 23:26, 8 February 2020 (UTC)

@Miraclepine: we wouldn't generally act on requests such as these. As we are accountable for additions we need to know what we are managing. I would suggest that you take it to enWP's ANI and get them to look whether this is the best way to address whatever issue you are seeing.  — billinghurst sDrewth 13:51, 9 February 2020 (UTC)
@Billinghurst: The enwiki TB says that "<newaccountonly> is no longer useful locally as accounts are created globally, these would need to be added to the global blacklist." The LTA in question is WhenDatHotlineBling, and accounts associated with that user (and also another LTA at enwiki, My Royal Young) keep being created. ミラP 00:24, 10 February 2020 (UTC)
I understand the issue, and I did not say get them to blacklist them locally, please reread what I did say. For LTAs your suggested method of response is not robust, so we will work cooperatively with local admins and they will supply us with information for areas where they need our assistance.  — billinghurst sDrewth 07:24, 10 February 2020 (UTC)

デーモン閣下

To prevent any acts of creating a new account, please add the it to the blacklist due to it's long-term pattern of abuse,

.*デーモン閣下.<newaccountonly>

The many times the vandal began creating the デーモン閣下 ends up with the number or デーモン閣下XX. It needs to be stopped creating the range of LTA. 2600:6C4E:580:A:2CD7:FF26:B5F7:824 17:16, 15 February 2020 (UTC)

X mark.svg Not done I had this in mind too, but after some deliberations as well as consultation with ~riley, we had decided to monitor rather than adding into the title blacklist. No prejudice for addition in the future should it go out of hand. Thanks.--Camouflaged Mirage (talk) 17:19, 15 February 2020 (UTC)
Second this closure. If we title blacklist it, they will just change the username to a pattern we aren't monitoring. Let's wait for it to get worse. ~riley (talk) 17:33, 15 February 2020 (UTC)
@~riley and Camouflaged Mirage: Okay can you get Rxy to take a look on it's デーモン閣下 ranges ending with XX number, because I found デーモン閣下42 to デーモン閣下47, the remaining users which is blocked on jaWP. Can you prefer locking the remaining accounts now? 2600:6C4E:580:A:2CD7:FF26:B5F7:824 17:59, 15 February 2020 (UTC)
I had submitted a SRG request for locking of those accounts. To be honest I don't mind blacklisting, but the concern that they may jump to another name and the increase in difficulty to track them down is holding me back. I am on this case from the very beginning, rest assured I will not let their abuse go out of hand. --Camouflaged Mirage (talk) 18:18, 15 February 2020 (UTC)
This guy previously don't use demon lord name, so do not blacklist this one, use abusefilter instead so it'll be easier to spot when this guy came back with same name demonlordxxx something. Still need to run check though, I thought he only have 87 account previously but apparently this guy already reach 100+.--AldnonymousBicara? 18:23, 15 February 2020 (UTC)

My username

For new account only as there are a couple of gross attacks against me. I could have edited myself but I don't think I know regex enough so can someone help just not to break anything, much appreciated :) Camouflaged Mirage (talk) 13:22, 2 April 2020 (UTC)

Added--Sakretsu (炸裂) 18:47, 2 April 2020 (UTC)
Hey Sakretsu, can you please log this change when you get a chance? Thanks man! ~riley (talk) 22:20, 2 April 2020 (UTC)
Right, done. Thank you for the reminder :-)--Sakretsu (炸裂) 00:22, 3 April 2020 (UTC)

oof.*

Can you please add oof.* <newaccountonly> to the title blacklist because I was doing a quick wiki search and realize that almost every username that starts with "oof" has made an unconstructive edit.

Awesome Aasim 22:08, 26 March 2020 (UTC)

X mark.svg Not done that you block a variation of a name isn't going to change either the behaviour of the editor or the likelihood that the edits will be good  — billinghurst sDrewth 05:16, 3 May 2020 (UTC)

Usernames longer than a number of characters

We had this problem in the past and we're experiencing it back again. There's an LTA creating accounts with long names such as MAJIDE YouTube no CM no JYOSEI SINEYO OOOOOOOOOOOOOOOO OOOOOOOOOOOOOOOO OOOOOOOOOO or OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO (more on this archive). I cannot think of any legit reason why we'd like to allow such long usernames, which I also suspect may break users on mobile displays or small screens. In all cases I remember this kind of long usernames have been vandals or LTAs. I'd say usernames longer than 25 characters should not be allowed, at least temporarily. Ping to Billinghurst. —MarcoAurelio (talk) 08:46, 25 May 2020 (UTC)

Have we got any stats on username length? For example, a count of usernames 20 or more characters in length, and how any of these are active, locked, blocked at a wiki? We can do something about usernames with excessive repeated character reasonably easy. @Martin Urbanec: can you use your access to tell us the story of what we might be preventing in the real world. I am concerned that we are just going to end up with vandal names that are shorter, which are a little harder to spot, but no less an issue.  — billinghurst sDrewth 09:24, 25 May 2020 (UTC)
FWIW MediaWiki uses mw:Manual:$wgMaxNameChars to prevent usernames longer than 255 bytes by default. On Wikimedia we use a reduced 85 bytes value; but I'd be cautious reducing that value as it can make usernames 'invalid' as per docs. According to mw:Account creation user experience/Usernames enwiki restricted usernames longer than 40 chars. Your concerns are exactly why I decided to open this discussion first. Thanks. —MarcoAurelio (talk) 09:30, 25 May 2020 (UTC)

Title blacklist/Log/2009/04

Hey, would you get rid of that spam blacklist category, because it doesn't belong to the spam blacklists at all. 60.142.240.26 00:01, 29 May 2020 (UTC)

Black check.svg Already done  — billinghurst sDrewth 11:10, 29 May 2020 (UTC)

Maximum number of characters for a username

The limit should be 75.

I calculated this by using [1], and entered the signature into the field. It looks like this:

[[User:abcdefghijklmnopqrstuvwxyzabcdefghijklmnopqrstuvwxyzabcdefghijklmnopqrstuvwxyz|abcdefghijklmnopqrstuvwxyzabcdefghijklmnopqrstuvwxyzabcdefghijklmnopqrstuvwxyz]] ([[User talk:abcdefghijklmnopqrstuvwxyzabcdefghijklmnopqrstuvwxyzabcdefghijklmnopqrstuvwxyz|talk]])

Then, I removed the end characters to make sure the signature is under the limit of 255 characters. --Keyacom (talk) 19:59, 29 May 2020 (UTC)

@Keyacom: you can always use the w:en:Help:Pipe trick in that box, and you certainly can link with smaller labels. (FWIW I do think that a 75 characater username is way to long to be useful). — xaosflux Talk 15:51, 2 June 2020 (UTC)
I know, right? - Balloony --Keyacomsomething to say? 17:24, 2 June 2020 (UTC)

Request for adding some words on Serbian to blacklist

Hello, please add these words in blacklist (Serbian strong language):

  • геј (it means gay, also add word below which is on Latin)
  • gej
  • админ (it is admin)
  • admin (same word but on Latin)
  • бирократа (it is bureaucrat)
  • birokrata (same word but on Latin)
  • penis (We have in blacklist same word but on Cyrillic, latin should be added also)
  • (Ј|ј)еб* (it means fu**ing, let's add it like this, so this can be catched by regex correctly and accounts like this can't be registered, like this)
  • (J|j)eb* (Serbian Latin for previous)
  • (К|к)ур* (the most commonly used name on Serbian for the male you know what)
  • (K|k)ur* (Serbian Latin for previous)
  • (Л|л)езб* (Regex to prevent creating accounts which contains лезбејка and simular. It means same thing as lesbian on English)
  • (L|l)ezb* (Serbian Latin for previous)
  • (Ц|ц)ар* (Regex to prevent creating accounts which contains Цар, царина and simular. It means same thing as emperor on English)
  • (C|c)ар* (Serbian Latin for previous)

This is everything for now, I will create new later if I find something that should be added also. You can fix regex to work, if my is incorrect. Best regards, Zoranzoki21 (talk) 07:11, 26 May 2020 (UTC)

@Zoranzoki21: I cannot add these words to title blacklist just because they are strong language. I also cannot add a string like "penis" to the blacklist. I would get myself crucified. What are you trying to manage? Please remember title blacklist covers all the Serbian sister sites, and these are matched as strings.  — billinghurst sDrewth 08:43, 8 June 2020 (UTC)

Business Lessons

Hey, this title has been a target for spambots lately. Would you please add this entry to the title blacklist? I'll appreciate it! 27.117.207.14 01:56, 8 June 2020 (UTC)

X mark.svg Not done please point to examples, and it will be reviewed to the best way to handle the issue.  — billinghurst sDrewth 08:36, 8 June 2020 (UTC)

Pages and usernames ending with semicolon

I can edit the blacklist myself, but I've never done this and I wanted to get a code review first, so to speak. The issue is phab:T238285. Pages ending in semicolons are essentially inaccessible. I think we should disallow them (and usernames) until the bug is fixed. I was going to use this:

# Until T238285 is fixed
.*;$

Does that look correct? Will this disallow both page titles and usernames, or do they need to be separate rules? MusikAnimal talk 22:37, 8 June 2020 (UTC)

Not certain we need the terminating $ as the instruction says pre and post wildcards. As we cannot see any logs for title blacklist, I suggest do one of the patterns, and test.  — billinghurst sDrewth 04:05, 9 June 2020 (UTC)
Yes check.svg Done Appears to be working! I did use $ as I see it's used in one other rule. It might not be necessary, though. Thanks for the help! MusikAnimal talk 04:46, 9 June 2020 (UTC)
Yeah, it works but the terminating $ isn't needed. It's as if you wrote ^.*;$$--Sakretsu (炸裂) 11:52, 9 June 2020 (UTC)

Circled and parenthesized Latin letters

Should be disallowed for new accounts. I reckon User:ⒶⓝⓣⓘⒶⓓⓜⓘⓝ ①④\⑧⑧ should not have been allowed to be created on enwiki.

The proposed addition is: .*[\x{249C}-\x{24E9}].* <casesensitive | newaccountonly> # Circled and parenthesized Latin letters (slightly modified from w:en:MediaWiki:Titleblacklist).

Best, --Mdaniels5757 (talk) 16:54, 8 June 2020 (UTC)

@Mdaniels5757: There is a whole amount of unicode letters that have become problematic due to LTAs, and there is a phabricator ticket relating to these and their antispoofing (account and article names). It is my belief that to block swathes of these unicode letters requires a specific community consensus rather than general comment and requests for partial ranges on this talk page. If the admin corp of enWP was pass a resolution then that would be a means to bring a proposal here or to a phabricator ticket. I would much prefer that the exclusion be done at a deeper level than title blacklist if we are blocking swathes of unicode for account names, or article names.  — billinghurst sDrewth 22:20, 10 June 2020 (UTC)

Usernames with "="

The current template system doesn't work well with an equal sign. And if it can be justified for some pages, it creates a lot of problems in usernames. — putnik 16:23, 26 May 2020 (UTC)

X mark.svg Not done @Putnik: Use a positional parameter. Discussion has been had previously, and not applied. That someone doesn't write good templates, or instructions, is not a reason to exclude a character. If truly problematic, such a change should be made within the Mediawiki system, not through the title blacklist.  — billinghurst sDrewth 11:14, 29 May 2020 (UTC)
If there was any discussion about this, please indicate where it was.Carn (talk) 13:25, 30 May 2020 (UTC)
@Carn: You can read and search the archives as well as I can.  — billinghurst sDrewth 13:28, 30 May 2020 (UTC)
But unlike you, I do not claim that such events took place and I can’t remember the keywords for searching. If you were just silent, it would have been more polite. This question is interesting to me only as a solution to a problem for which you refused to do what was proposed as a solution. Carn (talk) 14:13, 30 May 2020 (UTC)
I am fairly certain that you can read the archives as well as I can, and the conversation is there. If you don't like attitude, then please don't bring it. It is not apparent that you even looked at the archives prior to issuing your challenge.

The solution, if required, is through system configuration, not an ugly hack through the title blacklist, and of course I refused to do it. Experienced admins don't do ugly hacks that just load up the system. Experienced admins don't apply hacks that have no retrospective effect, so would not resolve your current issue of your not having a ping that works.  — billinghurst sDrewth 14:36, 30 May 2020 (UTC)

Just passing by. It looks this could also be achieved with mw:Manual:$wgInvalidUsernameCharacters; but I am not sure such a request would be granted. —MarcoAurelio (talk) 16:08, 30 May 2020 (UTC)
Yes, I'm sorry, I didn’t catch the archives of this page, but using them now I saw that this question popped up only twice: Talk:Title blacklist/Archives/2018#= and Talk:Title blacklist/Archives/2015#Equals sign, and I would not describe what was happening on the links as discussions. Rather, you, as in this case, explained why you will not do this. I’m writing not for an answer, but simply to fix the state of the situation and attach related links. Carn (talk) 14:01, 2 June 2020 (UTC)
I do not think that your boorish attitude is at least somehow justified.It would be great if you will be polite. Iniquity (talk) 18:33, 18 June 2020 (UTC)

Falun Gong and LTA

Proposed to add the following:

.*法[輪轮](大法|功).* <newaccountonly>

This should not be allowed in user name (likely promotional) and is recently abused by LTA ([2]).--GZWDer (talk) 16:15, 20 June 2020 (UTC)

Yes check.svg Done addition  — billinghurst sDrewth 23:38, 21 June 2020 (UTC)

g. prinzi

I suggest removal of

.*giovanni.{0,5}prinzi.*
.*prinzi.{0,5}giovanni.*
.*giuseppe.{0,5}prinzi.*
.*prinzi.{0,5}giuseppe.*

These entries seems to be outdated (request entry), and they caused an obstacle for a new user who created the unrelated de:Giuseppe Prinzivalli (reported on (german admins noticeboard)). --MBq (talk) 15:44, 24 June 2020 (UTC)

I remove the entry that affects this article, there are other crosswiki articles that would be affected if the article did not exist yet. As I do not have a precise of the reason for the entry, I leave the others, someone who knows the reason. -𝐖𝐢𝐤𝐢𝐁𝐚𝐲𝐞𝐫 👤💬 16:08, 24 June 2020 (UTC)

The other names should also be removed, these also entered illogically, please:

  • .*giovanni.{0,5}prinzi.*
  • .*prinzi.{0,5}giovanni.*

--5.171.45.155 16:38, 24 June 2020 (UTC)

additional requests X mark.svg Not done at this time; rationale required.

@MBq: These have been added by a steward due to years of abuse by a user. They should only be removed in consultation, especially with that steward. One cannot say that they are outdated, as how would you know? These vanity spammers keep at things for years, their vanity does not go away. Where there are specific and immediate issues for your wiki, then your local "mediawiki:titlewhitelist" are a good place to start.  — billinghurst sDrewth 23:47, 24 June 2020 (UTC)

The regex "*.*giovanni.{0,5}prinzi.*" should not affect anyone, at least there are no articles in any project.

@WikiBayer: It is not so. The regex "*.*giovanni.{0,5}prinzi.*" unjustly hinders other wiki projects, like these ca:Giovanni Prinzi, and lb:Giovanni Prinzi. For this reason I would say that this regex should be removed --5.171.207.203 10:23, 25 June 2020 (UTC)

I found these pages myself, as you can see from my request for deletion.---𝐖𝐢𝐤𝐢𝐁𝐚𝐲𝐞𝐫 👤💬 10:44, 25 June 2020 (UTC)

@WikiBayer: So, the regex "*.*giovanni.{0,5}prinzi.*" can't be removed? --5.171.207.203 10:53, 25 June 2020 (UTC)

IP: Stop your trolling. And participate constructively As long as there is no sufficient reason and use for the removal, no admin will remove this regex.--𝐖𝐢𝐤𝐢𝐁𝐚𝐲𝐞𝐫 👤💬 11:03, 25 June 2020 (UTC)

On wheels

[3] this should probably have the "moveonly" option or only block non-standard versions, as English Wikipedia has 64 articles, a disambiguation page, and at least 16 redirects with "on wheels" in the title, and several others that could be created or moved to titles that include this (as well as the possibility of talk pages and archives, featured article nominations, and requests for deletion for any of these articles). Peter James (talk) 22:54, 1 July 2020 (UTC)

@Peter James: Probably best to get enWP to do some whitelisting through their local w:en:mediawiki:titlewhitelist. This vandal xwiki vandalises, and these other languages are unlikely to need that phrase.  — billinghurst sDrewth 11:18, 2 July 2020 (UTC)
Which I have done.  — billinghurst sDrewth 11:33, 2 July 2020 (UTC)

$128餐蛋飯

Proposed to add the following:

.*$128餐蛋[飯饭].* <newaccountonly>

This should not be allowed in user name (likely promotional) and is recently abused by LTA ([4]).--12З4567 (talk) 05:25, 3 July 2020 (UTC)

X mark.svg Not done LTAs will do that and then do variations. Title blacklist is not the way to stop the abuse, it just stops a certain variation, which is not what we are trying to do. Happy to hear from the admins of your community about issues that we can have to look at.  — billinghurst sDrewth 06:32, 3 July 2020 (UTC)

Marian Richero

The following is prone to false positives:

.*(richero|marian).*(maria|richero).*

An example is "File:Joannis Marianae Hispani, e Societate Jesu, De rege et regis institutione libri III - ad Philippum III Hispaniae regem catholicum. Eiusdem De ponderibus & mensuris liber (IA joannismarianaeh00mari).pdf". There are two main problems:

  1. It matches on "marian ... maria" as well as "richero ... richero", which I'm sure is not the intent of the filter. Independently piping each portion is a bad idea.
  2. I don't think the mere existence of "richero" and "maria" in a title are sufficient to get it banned. They should at least have some proximity.

I propose instead:

.*(marian[^a-z]*richero|richero[^a-z]*marian).*

King of ♥ 16:34, 30 June 2020 (UTC)

@King of Hearts: Yes check.svg Done I have just done it as two lines. It all gets turned into a mega-regex anyway.  — billinghurst sDrewth 16:21, 5 July 2020 (UTC)
@Billinghurst: I think you have an unclosed paren. -- King of ♥ 21:31, 5 July 2020 (UTC)
👍  — billinghurst sDrewth 06:02, 6 July 2020 (UTC)

Taurino Araújo

See cross-wiki spam. Ping reporter @Stanglavine:. As the subject is creating articles about himself on various wiki language variants, an entry on the blacklist may help. — Alexis Jazz (talk or ping me) 15:12, 5 July 2020 (UTC)

While this remains a suitable topic at some wikis,it seems premature for title blacklist at this point. Where it is clear spam it would be acceptable, however, there are claims of notability that are not the job of this list to make without a clear community consensus.  — billinghurst sDrewth 16:13, 5 July 2020 (UTC)
An entry on the title blacklist wouldn't prohibit any wiki from having an article, it would just require a local user with tboverride to create the page or move a draft to its final location. Also, the article about Araújo has been nominated/tagged for deletion on all Wikipedias. The burden of the spam seems greater than the cost of needing a user with tboverride on the few (if any) wikis that actually want an article about him. Also note that on Dutch Wikipedia (and so probably on others), the article was utterly worthless machine translation garbage. — Alexis Jazz (talk or ping me) 20:21, 5 July 2020 (UTC)
Not that I oppose consensus. How/where could that discussion best be held? (could it simply be here?) — Alexis Jazz (talk or ping me) 20:25, 5 July 2020 (UTC)

14 articles have already been deleted in several wikis and, of the remaining 8, all are tagged for deletion. Of these, 7 were created by two accounts or proxies (which is already an element that raises many suspicions about the intention). As an intermediate solution, I suggest we wait for the remaining wikis to decide on the elimination, to dispel any doubts about the notability. We can keep the discussion on hold for now, but I already say that I am in favor of inclusion on the blacklist. Rafael (stanglavine) msg 22:43, 5 July 2020 (UTC)

Domain name instead of username

The expression for determining the domains in the username is incorrect, because it not covers all TLD. For example, .ru, .io, .cn etc. Full list is here: https://www.iana.org/domains/root/db

# domain names . * [\. ] (?: com | org | uk | net | info | gov | kz | ru | ir | biz | info | short | pt | br) \ b. * <newaccountonly> --Ctac (talk) 17:22, 7 July 2020 (UTC)

@Ctac: I am not certain that it has ever been considered a proactive list, typically we are only blacklisting reactively to problems. If you are seeing TLD domain names being used that are problematic, then please suggest those that we should add. As a note, due to the lack of logging on title blacklists we typically try a cautious approach to additions.  — billinghurst sDrewth 00:18, 8 July 2020 (UTC)

Taurino Araújo

See cross-wiki spam. Ping reporter @Stanglavine:. As the subject is creating articles about himself on various wiki language variants, an entry on the blacklist may help. — Alexis Jazz (talk or ping me) 15:12, 5 July 2020 (UTC)

While this remains a suitable topic at some wikis,it seems premature for title blacklist at this point. Where it is clear spam it would be acceptable, however, there are claims of notability that are not the job of this list to make without a clear community consensus.  — billinghurst sDrewth 16:13, 5 July 2020 (UTC)
An entry on the title blacklist wouldn't prohibit any wiki from having an article, it would just require a local user with tboverride to create the page or move a draft to its final location. Also, the article about Araújo has been nominated/tagged for deletion on all Wikipedias. The burden of the spam seems greater than the cost of needing a user with tboverride on the few (if any) wikis that actually want an article about him. Also note that on Dutch Wikipedia (and so probably on others), the article was utterly worthless machine translation garbage. — Alexis Jazz (talk or ping me) 20:21, 5 July 2020 (UTC)
Not that I oppose consensus. How/where could that discussion best be held? (could it simply be here?) — Alexis Jazz (talk or ping me) 20:25, 5 July 2020 (UTC)

14 articles have already been deleted in several wikis and, of the remaining 8, all are tagged for deletion. Of these, 7 were created by two accounts or proxies (which is already an element that raises many suspicions about the intention). As an intermediate solution, I suggest we wait for the remaining wikis to decide on the elimination, to dispel any doubts about the notability. We can keep the discussion on hold for now, but I already say that I am in favor of inclusion on the blacklist. Rafael (stanglavine) msg 22:43, 5 July 2020 (UTC)

Tested positive

I propose modifying the line .*tests? positive.* to .*test(s|ed)? positive.* as an obvious workaround. LaundryPizza03 (talk) 15:04, 25 August 2020 (UTC)

@LaundryPizza03: What battle are we facing? Can you get support of your local administrators to support a change.  — billinghurst sDrewth 15:09, 25 August 2020 (UTC)
@LaundryPizza03: Final ping. Camouflaged Mirage (talk) 13:27, 4 September 2020 (UTC)
X mark.svg Not done No reply. Camouflaged Mirage (talk) 13:13, 6 September 2020 (UTC)
This section was archived on a request by: Camouflaged Mirage (talk) 13:13, 6 September 2020 (UTC)

Fix pseudo IPs rule

Despite:

# fake ipv4
.*(?:[0-9]{1,3}\.){3}[0-9]{1,3}.*        <newaccountonly>
 [^:]+:(?:[OoIl0-9]{1,3}\.){3}.*         <newaccountonly> # no pseudo IP adresses

I have recently locked Special:CentralAuth/57.124.З2.45 and some more following that pattern. We should amend the rule to prevent this from happening. Ping @Lustiger seth and Billinghurst. Thanks, —MarcoAurelio (talk) 09:47, 30 August 2020 (UTC)

hi!
for digits [0-9] we could use [0-9\x{ff10}-\x{ff19}OΟОՕOoοо༠oIlշЗ༢༣༥Ց৪] (but there are still probably many more)
furthermore: what about ipv6? there we should add [A-Fa-fАВСЕасе] and many more.[5] -- seth (talk) 11:06, 30 August 2020 (UTC)
This aligns with Talk:Title blacklist/Archives/2020#Circled and parenthesized Latin letters. MA I will send you an IRC memo.  — billinghurst sDrewth 11:50, 30 August 2020 (UTC)
the linked discussion is far more general. i don't think that we need an additional community consensus for blocking ip-address-like names.
extending the existing TBL entries could be a reasonable short- and mid-term solution. a long-term solution could be the modification of the mediawiki extension. -- seth (talk) 13:47, 30 August 2020 (UTC)
@Billinghurst @Lustiger seth: I'm not sure why .*(?:[0-9]{1,3}\.){3}[0-9]{1,3}.* didn't catch 57.124.З2.45 or 82.141.З5.28. Per https://regexr.com/5b4v0 it seems to work for every other IP. Is there anything I am not seeing? —MarcoAurelio (talk) 09:38, 31 August 2020 (UTC)
@MarcoAurelio: 57.124.З2.45  — billinghurst sDrewth 15:17, 31 August 2020 (UTC)
hi user:MarcoAurelio! (edit conflict)
echo 57.124.З2.45 | hexdump -C gives:
00000000 35 37 2e 31 32 34 2e d0 97 32 2e 34 35 0a |57.124...2.45.|
this means: "З" is not a three, but a en:Ze_(Cyrillic). same with "82.141.З5.28".
that's why i made the proposal of [0-9\x{ff10}-\x{ff19}OΟОՕOoοо༠oIlշЗ༢༣༥Ց৪] -- seth (talk) 15:25, 31 August 2020 (UTC)
Oh, very subtle, hardly visible to me. Thanks for the answer, both. —MarcoAurelio (talk) 15:29, 31 August 2020 (UTC)
at least for the ipv4 pseudo addresses i changed the entries now. -- seth (talk) 09:03, 1 September 2020 (UTC)

Remove caseless modifiers

Since TitleBlacklist extension is case insensitive by default, I propose to simplify the regex patterns as done in [6]. Dalba 02:56, 8 August 2020 (UTC)

I don't feel competent to be able to comment, I would suggest that you get the support of native speakers who are administrators to pass comment.  — billinghurst sDrewth 15:07, 25 August 2020 (UTC)
X mark.svg Not done (at least for now) while this could be discussed more, some robust multi-alphabetic test cases should be reviewed first. — xaosflux Talk 13:24, 4 September 2020 (UTC)

Obscure letter lookalikes and funny characters in usernames

.*[\x{2100}-\x{214F}].* <casesensitive|newaccountonly> # Letterlike Symbols
.*[\x{2460}-\x{24FF}].* <casesensitive|newaccountonly> # Enclosed Alphanumerics
.*[\x{2500}-\x{259F}].* <casesensitive|newaccountonly> # Box Drawing and Block Elements characters
.*[\x{1D00}-\x{1DBF}].* <casesensitive|newaccountonly> # Phonetic Extensions + Supplement
.*[^\0-\x{FFFF}].* <casesensitive|newaccountonly> # Very few characters outside of the Basic Multilingual Plane are likely to be used by good-faith accounts.
(Non-printing characters)

Rationale: Potentially deceptive usernames, unlikely to be used in good faith. See also en:MediaWiki_talk:Titleblacklist#Potentially_deceptive_characters_in_usernames. LaundryPizza03 (talk) 18:00, 7 September 2020 (UTC)

Declined Declined at this time. I do not think that listing every unicode set is the way to manage this situation. If this is to progress it would be better to occur through systematic approach of the underlying mediawiki coding. Use of title blacklist is an ugly sledgehammer and, as we cannot see logs, it just makes things uglier than needs to exist. Please lodge a phabricator ticket if you wish to progress this matter.  — billinghurst sDrewth 22:13, 7 September 2020 (UTC)
@LaundryPizza03: what standing are you bringing to this conversation. I would encourage you to engage with administrators at your local wiki as the initial means to discuss whatever problem you are facing. Those administrators should escalate as required.  — billinghurst sDrewth 22:15, 7 September 2020 (UTC)

enwiki admin here. I agree that blocklisting everything outside the BMP is a bad idea; it's just too broad an exception, and covers cases of exotic but valid writing systems that should not be blocked. However, blocking the Letterlike Symbols and Mathematical Alphanumeric Symbols, as well as the Phonetic Extensions and supplement, for username creation alone, seems like a good idea to me, as they can be used to create usernames with decorative "font effects" that are both confusing to editors, and impossible to re-type or search for with ordinary text search. See, for some real-world examples, 𝗺𝘆 𝗺𝗼𝗺 𝗶𝘀 𝗽𝗿𝗼𝘂𝗱 𝗼𝗳 𝗺𝗲 and 𝕷𝖔𝖗𝖉 𝕺𝖋 𝖘𝖍𝖆𝖉𝖔𝖜𝖘 𝖝𝖊𝖓.

Because this is username creation, it can't be performed at the local wiki level, and needs to be performed globally. -- The Anome (talk) 11:19, 8 September 2020 (UTC)

.*[\x{1D400}-\x{1D7FF}].* # en:Mathematical Alphanumeric Symbols

The block also includes lookalikes of Greek letters and numerals. Are there other characters that are used for impersonation? LaundryPizza03 (talk) 14:34, 8 September 2020 (UTC)

Letterlike_Symbols, Phonetic Extensions and Phonetic Extensions Supplement can also be used, although with more effort, since they have a much smaller repertoire. All of these are for technical use only, and none of them are required to encode any commonly used writing system.

While I think there's a case for dealing similarly with the enclosed alphanumerics and box drawing characters, they don't have the same potential for spoofing as these. Let's just concentrate on getting these (I hope) relatively uncontroversial spoofing/"font-style" ranges through, and treat the other ranges of potentially-funny characters as a separate issue. -- The Anome (talk) 19:05, 8 September 2020 (UTC)

Aybars İbak

  • Aybars İbak
  • Aybars Han İbak
  • Aybars Ibak

Reasoning: repeated cross-wiki spam of this subject with ZERO notability that's been going on for over a year. Speedy deleted from over a dozen wikis, I just indefinitely protected it from creation on it.wiki. A research on Google shows nothing more than a self-directed short movie and a self-released song. A number of users (probably the same person with multiple accounts) is responsible for the relentless spammage. Merynancy (talk) 11:56, 21 September 2020 (UTC)

jaWP LTA username regex

For ja:LTA:ASPE

# Detect anagrams
^(?=.*a)(?=.*s)(?=.*p)(?=.*e.*e)(?=.*r.*r)(?=.*g).{8}$
^(?=.*ア)(?=.*ス)(?=.*ペ)(?=.*ル)(?=.*ガ)(?=.*ー).{6}$
.*Asperger.*
.*アスペ.* # (A abbreviation of アスペルガー)

w:ja:LTA:ASPE uses anagram of アスペルガー (It means Asperger in Japanese) in account names.--Semi-Brace (talk) 09:44, 24 October 2020 (UTC)

The 3rd one looks fine; not sure if the remaining ones will cause false positives. ネイ (talk) 13:35, 15 November 2020 (UTC)
@ネイ: I don't think 4th one causes false positive; see quarry:49890quarry:query/49890. (284 out of 296 are indef-blocked in jawp). --Semi-Brace (talk) 13:22, 17 November 2020 (UTC)
Also, 1st won't cause false-positives (in other words, won't match other than anagrams): test for 1st.
2nd one may cause some sacrifice (e.g. アールペガス). but, won't match other than anagrams: test for 2nd. —The preceding unsigned comment was added by Semi-Brace (talk) --Semi-Brace (talk) 09:33, 25 November 2020 (UTC)
For the 1st one I will defer it back to billinghurst; if the 2nd one causes sacrifices we better be more conservative on that; for the 4th one, the quarry link is not working so I cannot comment for now. ネイ (talk) 05:48, 25 November 2020 (UTC)
Fixed broken link. --Semi-Brace (talk) 09:33, 25 November 2020 (UTC)

For ja:LTA:NDBTK

.*タヒね.*

w:ja:LTA:NDBTK uses "タヒね" in account name, which is variant writing of "死ね" (means "die" in Japanese; it's strong language).--Semi-Brace (talk) 09:44, 24 October 2020 (UTC)

This looks fine to me. ネイ (talk) 13:35, 15 November 2020 (UTC)

For ja:LTA:IKE

.*(知的?障害?|池沼).* # current
.*(知的?障害?|池 ?沼).* # after change

w:ja:LTA:IKE uses "池 沼" in account name, which is variant writing of "池沼".--Semi-Brace (talk) 09:44, 24 October 2020 (UTC) PS --Semi-Brace (talk) 09:55, 24 October 2020 (UTC)

I prefer adding detection of whitespaces or symbols in between, which is less wide than a blanket .?. The widest change I can accept for now is to detect non-kanji/kana character in between. ネイ (talk) 13:35, 15 November 2020 (UTC)
@ネイ: replaced with " ?". --Semi-Brace (talk) 05:34, 17 November 2020 (UTC) ping --Semi-Brace (talk) 13:11, 17 November 2020 (UTC)
This one looks fine after the change. ネイ (talk) 14:52, 17 November 2020 (UTC)

comment

@Semi-Brace: Please get the support of the jaWP's administrators for these regex, and please ensure get the direction whether we are managing usernames and/or pages. If we are talking pages, then please also check these with administrators at places like jaWikt, jaWS, and other ja language communities that may be appropriate.  — billinghurst sDrewth 09:50, 24 October 2020 (UTC)

@Billinghurst: I'm sorry for ambiguity of propose. It doesn't target pages, but accounts.--Semi-Brace (talk) 09:55, 24 October 2020 (UTC)
Noted, and adjusted title. For these sorts of adjustments I would generally get both stewards and/or local admins to recommend progressing.  — billinghurst sDrewth 10:00, 24 October 2020 (UTC)
Commented on the changes above. Thanks. ネイ (talk) 13:35, 15 November 2020 (UTC)
@ネイ: Answered questions about posibility of false-positives from you, please check, thank you. --Semi-Brace (talk) 06:06, 23 November 2020 (UTC)
Replied. ネイ (talk) 05:48, 25 November 2020 (UTC)
@ネイ: Fixed broken link per above notice; see quarry:query/49890 for details of that. --Semi-Brace (talk) 09:33, 25 November 2020 (UTC) ping --Semi-Brace (talk) 13:28, 26 November 2020 (UTC)

Anan Islam

Please blacklist Anan Islam, Md Anan Islam and मोहम्मद आनान इस्लाम. This person have a long story of cross-wiki spam lasting more than six months and several dozens of sockpuppets and global bans. w:Wikidata:Requests for checkuser/Case/Ananislam138. --Trade (talk) 20:09, 31 December 2020 (UTC)

@Trade: User name, or page name? Will it work, they typically just go for a variation outside of the listing. Have you checked global-search for collateral damage? I will add to the ego AF filter that is in place  — billinghurst sDrewth 01:19, 1 January 2021 (UTC)
Page name. They don't use their real name as account name anymore. --Trade (talk) 10:38, 1 January 2021 (UTC)

A false positive

Just to make a note of it, on English Wikisource we just had a false positive hit on the entry .*(mattia|vlad|morleo).*(mattia|vlad|morleo)?.*(mattia|vlad|morleo).*, for the page s:The Conversion of St. Vladimir/A Short Historical Sketch of the life of St. Vladimir (the subpage is the title of a chapter within a book whose title is the parent page name).

This is a likely but not frequent page name on enWS, and the regex looks to me like it attempts to avoid hitting probable legitimate cases, so I'm just noting it here as input to the assumptions underlying that. I don't think there's any particular need to change the blacklist entry (it's bit us just once in the ~2 years the entry's been there). --Xover (talk) 15:48, 7 September 2020 (UTC)

The regex construction is unusual, as it either needs two matches, or it needs three matches. It doesn't actually make sense to have the middle set to be conditional as that is just the continuation of the all characters wildcard. I would suggest that we make the three parts the regex.  — billinghurst sDrewth 20:58, 5 November 2020 (UTC)
@Vituzzu: Re this special:diff/19102572 edit, I am trying to understand the benefit or the purpose of the second conditional match.  — billinghurst sDrewth 22:58, 5 November 2020 (UTC)
@Billinghurst: AFAIR both him and Giovanni Prinzi (the LTA pushing any name I put on title blacklist) use many different variations of this name: "mattia morleo", "morleo mattia", "vlad morleo", "morleo vlad", "mattia vlad morleo", "morleo mattia vlad", etc. --Vituzzu (talk) 14:45, 13 November 2020 (UTC)
@Vituzzu: That is not my question, this is not about the individual phrases. The regex has the same phrase three times, though makes the second optional, so in essence just is requiring two matches, not the three (we have redundancy). Two matches is showing false positives. So do we go back to three, or do we need to make the the middle section different so it skips titles with subpages?  — billinghurst sDrewth 20:26, 13 November 2020 (UTC)
If we need three then .*(mattia|vlad|morleo).*(mattia|vlad|morleo).*(mattia|vlad|morleo).*, whereas if we are wanting two and something to avoid Xover's example then we could user .*(mattia|vlad|morleo)[^\/]*(mattia|vlad|morleo).* which ignores the two when there is a forward slash.  — billinghurst sDrewth 20:50, 13 November 2020 (UTC)
Mine was indeed wrong, but I was wondering why did I write it that way. Anyway, as usual, you can always change my addition on need. --Vituzzu (talk) 21:38, 22 November 2020 (UTC)

Yes check.svg Done changing regex per discussion.  — billinghurst sDrewth 21:08, 18 January 2021 (UTC)

This section was archived on a request by:  — billinghurst sDrewth 21:08, 18 January 2021 (UTC)