User talk:Dispenser/Archive 5

Dabfix changes on songs

Hi, I just ran dabfix on Secrets and it mad a bit of a mess on the songs section. I saved the output to User:Tassedethe/Secrets. I'm not sure I'm too keen on the style for songs. It changed everything from "a song by" to "a song on". One problem is that getting the apostrophe for the artist/band could be tricky. I changed The Pierces's to The Pierces'. It also looks like a dabfix too far, there's nothing really wrong with how it was, although having it impose consistency could be good on some pages. But I'll still be using it, it's still much too useful to give up! :) While I'm unloading I think you changed the behaviour for how it deals with redlinks in templates. I agree that just because it has 10-20-whatever links it shouldn't be an automatic choice to go on the page but now it doesn't show up as existing. It would be useful if it did, it's quite common for the template link to be a match for the dab page but someone has used a non-standard disambiguator, or (if it is a name) there is actually a middle initial. Just letting you know my thoughts. Happy editing. Tassedethe (talk) 02:50, 6 January 2011 (UTC)

:I just added that today and now disabled it. The idea was to shorten as much as possible, mostly for people at MOS:DAB and I had been doing it manually myself. The apostrophe -s rules depend if are treating the band's name as singular (actress's) or plural (actresses'); I inclined to treat them as singular as there is only one band with that name. Then there is the problem of groups like Racionais MC's and Musashi's which it just skips.

:It is not trivial to determine if redlinks on a particular page are from a template. The red link might only appear on /doc and transcluded to the template page or can be hidden by <includeonly> tag. This screws up one of the more important heuristics. Note, that being listed on a WikiProject page or a Missing articles page will still force the link to show up. Additionally, dabfix hides a lot of extraneous detail behind debug (use "show debug" link) which is also where these template link vanish too. — Dispenser 06:46, 6 January 2011 (UTC)

::Thanks for the info. I thought perhaps the templates were something you'd disabled; obviously far more complicated than that. I will check the show_debug link in future. Ta. Tassedethe (talk) 14:44, 6 January 2011 (UTC)

Other dabfix problems

I've been liking the tool so far, but I've noticed a few issues. Particularly, the script seems to screw up with apostrophes, so that it gives phrases like "the band, 's album". It also seems to remove the italics from ship names as can be seen on Marigold. Today I also noticed that it seems to have issues with Japanese characters, as seen when trying to run the script on Bottom. As always, thanks for creating and maintaining another useful tool. —Ost (talk) 21:25, 12 January 2011 (UTC)

:I'm not sure what changed, but the problem on Bottom did not occur when I ran the tool again. —Ost (talk) 21:26, 12 January 2011 (UTC)

[[Worldwide]]

Not a big deal, certainly, but I was curious what the difference between {{tlc|Wiktionary}} and {{tlc|Wiktionary|worldwide}} was in this particular case. I only removed the "worldwide" parameter because I was already editing the page, and according to the template docs it's superfluous -- my "-dicdef" summary was in reference to removing the dictionary definition of "worldwide" from the body of the page.

Also no big deal on the date for the album either way, btw; you probably noticed I'd even left the date for the other album in. Cheers, NapoliRoma (talk) 14:51, 19 January 2011 (UTC)

: If the parameter is unspecified the template will pass it through the search function at Wiktionary (Special:Search/worldwide). My tool, Dab solver, will use the search page link instead of the direct page on Wiktionary when people try to link to the definition. — Dispenser 19:34, 19 January 2011 (UTC)

:: OK, but doesn't that veer into WP:DWAP territory?

:: Given that both forms take the reader to the same destination, it seems like we should be more worried about the maintainability of the article rather than what's happening underneath the hood today (which may be entirely different from what's happening underneath the hood tomorrow).

::If this is really a big performance hit, I think that at the least, the documentation for {{tl|Wiktionary}} should be updated to warn about that, or even better, some optimization in the back end was done to make it irrelevant to editors (which WP:DWAP seems to say it really is, even now).

:: I'm not entirely certain what you're saying about Dab solver, and how this enters into the discussion. Clarification? Regards, NapoliRoma (talk) 21:50, 19 January 2011 (UTC)

::: No, there's a semantic difference. The search implies least one page, possibly more (due to capitalization), while the defined form indicates only that/those pages. — Dispenser 00:06, 8 January 2012 (UTC)

Dabfix problems

:''migrated from User talk:Logan#Dabfix problems

The tool removed the wiktionary link because there was no Hot mess (disambiguation) (or similar qualifier) to the page. Thus, there is no a path for a reader reaching the disambiguation page to expect "hot mess" as a definition. Primary topic usage around it complicates it more so. However, I'll look into to add a Wiktionary verification in the future. Secondly point, heading the are more semantic then bold text and I'm using this to enhance the presentation in another tool. Use level 3 headings can be used if you visually like them better. — Dispenser 00:01, 21 January 2011 (UTC)

:Readers entering "hot mess" in the search box will be taken to Hot Mess and then, for other uses, to Hot Mess (disambiguation) -- I failed to save that change to the hatnote on Hot Mess, so perhaps now that it's there, the tool would recognize that? Level 3 headings without level 2s would not be an improvement. The non-level groups (like "In music:) are preferred over sections in short dab lists as they keep the list easier to use for the reader; the other tool may not have that issue, if the other style guide doesn't have that preference. Thanks! -- JHunterJ (talk) 12:12, 21 January 2011 (UTC)

Barnstar

style="border: 1px solid gray; background-color: #fdffe7;"

|rowspan="2" valign="middle" | 100px

|rowspan="2" |

|style="font-size: x-large; padding: 0; vertical-align: middle; height: 1.1em;" | The Barnstar of Diligence

style="vertical-align: middle; border-top: 1px solid gray;" | This is for your continued work on all of your Toolserver tools. Keep it up! Logan Talk Contributions 19:59, 22 January 2011 (UTC)

Report error?

This report, in the third row, shows Matthew Keating (disambiguation) as a redirect to Matt Keating. Actually, the title is a redirect to Matthew Keating, and the page history shows that it always has been. Seems like a possible bug in the script that generates the report. --R'n'B (call me Russ) 21:32, 24 January 2011 (UTC)

: For whatever reason, databases aren't always properly populated when they're save. Currently there's a script repopulating the redirect table for the new columns (rd_fragment and rd_interwiki) introduced with bug 14418. (It may have died actually) — Dispenser 02:58, 17 January 2012 (UTC)

Wayback Machine

Drop by to tell you new beta is up, http://waybackmachine.org/ « ₣M₣ » 22:47, 25 January 2011 (UTC)

:And I dropped by to let you know that I think that the Checklinks code needs to be updated to use the new site; clicking "Internet Archive" from the tool gives a 404 Page Not Found error. Thanks as always. —Ost (talk) 18:15, 5 July 2011 (UTC)

::Unfortunately, the code I wrote five years ago wasn't designed well, especially with data structures. A complete rethink is required so separations are clean and well defined. This includes rethinking the user interface so its simple and discoverable. I'm currently working to cleanup the periphery stuff, so its going to be a while till I get there. — Dispenser 03:53, 17 January 2012 (UTC)

WikiProject activity check tool

Hi, your activity check tool has been brought up in this discussion at the WikiProject Council regarding WikiProject activity. I think that if the tool could be modified to only include a subset of pages (for example articles tagged with a WikiProject banner), then it could be used to measure the activity / inactivity of entire WikiProjects. Do you think this is possible? Thanks a lot, Mlm42 (talk) 18:17, 26 January 2011 (UTC)

: I've actual done some work on this many time. Having a large article collection allows for some interesting things, like a group watcher, hours spent editing, gender ratio, group watchlist/activity list, geolocate IP editors, disambiguation challenges, and more. I have dabbled in all of these, but to integrate them well takes more time then one would suspect. I am still interested, but need to be shown its worth the time invested. — Dispenser 21:30, 14 August 2011 (UTC)

Test your Simpsons knowledge

Brilliant. --JaGatalk 05:50, 7 February 2011 (UTC)

coord-dewiki.log

Hi Dispenser, why is the region "RU-ZAB" listedd as error in coord-dewiki.log ? --91.22.202.234 (talk) 09:21, 14 February 2011 (UTC)

: Unfortunately, a reasonable complete and up to date (last 5 year) machine readable list is not freely available. Thus, I've removed manually inputted RU-* subdivision codes. Someday I'll have to scrape the information off Wikipedia and make it available to everyone. — Dispenser 23:24, 27 February 2011 (UTC)

Rdcheck bug

This might be a toolserver bug, not a rdcheck bug. But it would seem that anytime I click a link on rdcheck with a link with an ampersand (&) the link does not work.

see this: http://toolserver.org/~dispenser/cgi-bin/rdcheck.py?page=Texas_A%26M_University

Now, if I were to click on a link like Texas A&M, i would only get Texas A. See http://en.wikipedia.org/w/index.php?title=Texas%20A&M&redirect=no

A similar bug was found in other tools, so I think it might be a toolserver thing. thanks for your help. Oldag07 (talk) 02:30, 15 February 2011 (UTC)

coor_g.php

Any progress in getting me added as a maintainer of geohack? I am now confident that I can add the stuff myself. The modules and changes involved are listed here. You might care to do a quick check of my version of geohack.php. — RHaworth {{toolbar|separator=dot|talk | contribs }} 17:31, 27 February 2011 (UTC)

:TS-959Dispenser 22:44, 27 February 2011 (UTC)

Love your work

The sound boxes at "F and A"—nice. Tony (talk) 07:51, 3 March 2011 (UTC)

Commonfixes

Hi. There might be an error with the tool. Please check the following links, and keep your view highlighted on the swingout boxes in the article

(div style="clear:both; class="NavFrame" changing into --->

  • http://de.wikipedia.org/w/index.php?title=Star_Wars:_The_Clone_Wars_%28Fernsehserie%29&diff=85990766&oldid=85965347
  • http://de.wikipedia.org/w/index.php?title=Star_Wars:_The_Clone_Wars_%28Fernsehserie%29&diff=next&oldid=85990766
  • http://de.wikipedia.org/w/index.php?title=Star_Wars:_The_Clone_Wars_%28Fernsehserie%29&diff=next&oldid=86004052

:Thx. -- Gary Dee (talk) 16:24, 3 March 2011 (UTC)

F and A format

Hi, thanks so much for organising the sounds last week. This time, I wanted to move the sounds up, possibly to be just under the admins. But I can't get rid of the header "Want the latest Signpost delivered to your talk page each week?" that intrudes just above it. Can you help? Link Tony (talk) 10:11, 5 March 2011 (UTC)

:Thanks; my computer IQ is –5. Tony (talk) 14:26, 5 March 2011 (UTC)

Single line sorting for wikEd

Please check User_talk:Cacycle/wikEd#Req:_Sort_.28and_normalize.3F.29_list_entries_contained_in_a_single_line, I have added single line sorting to wikEd. Sorting is alphabetically an corrected for numbers. Cacycle (talk) 21:45, 6 March 2011 (UTC)

coord elevation parameter

Several years back, the documentation for {{tl|coord}} and/or WP:GEO (possibly it was a WT:GEO discussion) described an elevation parameter, and went into some detail about units and how it improves on the type:mountain(nnnn) for other situations. I don't see the description anywhere now, but I've faithfully added elevations when doing other work to a coordinate. Unfortunately, your tool flags these which some people have "repaired" by deleting the parameter. Maybe this is an appropriate discussion for WP:GEO (or maybe it already occurred, but I missed it), but how do you feel about accepting elevation:? Awhile back I saw another digest of coordinates which recognizes it—maybe that was your work as well? —EncMstr (talk) 08:15, 10 March 2011 (UTC)

new values for the Coord globe: parameter

I was intrigued to see that you've been implementing new globes for GeoHack (such as Jupiter). Very cool! However, when you do this, you should document the change at Wikipedia:WikiProject Geographical coordinates/globe:. Also, I notice that the Jupiter test case you added to Template talk:GeoTemplate/test coordinates is broken; the red dot does not coincide with the Great Red Spot. —Stepheng3 (talk) 03:12, 13 March 2011 (UTC)

Enaturalist

Hi. Could you tell me please, why [http://www.enaturalist.org/biography.html this] domain is listed as a WP-Blacklist URL. See tools:~dispenser/cgi-bin/webchecklinks.py/de:William-Brewster-Medaille . Thx -- Gary Dee (talk) 13:05, 16 March 2011 (UTC)

Template thumbnail

Hi, I don't quite understand its purpose/function. The link to the example didn't work for me. Any ideas? Tony (talk) 02:08, 29 March 2011 (UTC)

Modern skin

Hello Dispenser, I was using Reflinks, and I noticed it was now adding &useskin=modern to the end of the URL when the preview or show changes button is used. Is there a setting that I can use so Monobook is automatically used, rather than removing &useskin=modern manually each time? Thanks, Alpha Quadrant talk 23:09, 1 April 2011 (UTC)

: Happy April Fools day :-). — Dispenser 00:58, 2 April 2011 (UTC)

:: Ha, nice one :). At first I had thought I was still on toolserver, before I realized the &useskin=modern. Very clever joke. Alpha Quadrant talk 01:12, 2 April 2011 (UTC)

:::Also, bug 22029 is for preserving useskin between edit screens. — Dispenser 02:13, 2 April 2011 (UTC)

Autoconformation RfC

A formal Request for Comment has now been started on this topic. Feel free to contribute; best, Ironholds (talk) 19:30, 3 April 2011 (UTC)

Your Geohack Google maps tool

Hi. I'm interested in using your Geohack tool [https://wiki.toolserver.org/view/GeoHack/Replacement_script here], but it doesn't seem to have any documentation or any explanation of what to do with all that lovely code. You should probably write a line or two in the commented-out section at the top as to where to put the code to actually use your script! Thanks, Oreo Priest talk 00:11, 7 April 2011 (UTC)

Bug with Rdcheck

If there is a "+" within a redirect lemma then it will be wrong encoded on the result page. So if a request like this one is started then you'll get invalid results like this:

  • http://de.wikipedia.org/w/index.php?title=U+2318&redirect=no

instead of this:

  • http://de.wikipedia.org/w/index.php?title=U%2B2318&redirect=no

--ζ 10:51, 20 April 2011 (UTC)

: Fixed (months ago, just documenting it). — Dispenser 21:21, 22 July 2011 (UTC)

Help to you

Hi Dispenser, i want to help for translate this [http://en.wikipedia.org/wiki/User_talk:Dispenser/Dab_solver#Translating_to_danish page] for the portuguese language. I want to know if i have just traslate that page, or in another place. Please, answer in wiki-pt page. Thanks. --Vitor Mazuco Talk! 15:47, 8 May 2011 (UTC)

Rdcheck and ancor templates

There seems to be an issue with Rdcheck and at least {{tl|visible anchor}}. It's coming back with invalid redirects when the template is involved. An example is here: tools:~dispenser/cgi-bin/rdcheck.py?page=List_of_characters_in_The_Batman

- J Greb (talk) 17:14, 23 May 2011 (UTC)

ghel database problem

:<class 'oursql.ProgrammingError'>: (1049, "Unknown database 'u_dispenser_p'", None)

Hi, it seems there's a problem with ghel's database, would you please take a look at the error?? Thanks! --Joana.estafanell (talk) 10:10, 2 June 2011 (UTC)

:Fixed The Toolserver admins have corrected the DNS entries so it points to the server containing user databases. — Dispenser 20:17, 2 June 2011 (UTC)

Request for new numbers of installs/active users of gadgets

Hi there,

I am the author of the GoogleTrans gadget. Thanks very much for those stats you put up on Sept 11 2009. It was good to get some feedback on the real use of the gadget. Is it a lot of trouble to get current stats like that for the gadgets? I'd certainly like to know how GoogleTrans is going right now.

Thanks Again

Endo999 (talk) 22:41, 3 June 2011 (UTC)

:Wikipedia:Database reports/User preferences, updates weekly. It's possible to build an adoption graph by collating the revisions. — Dispenser 01:08, 4 June 2011 (UTC)

Thanks for that. I just took a look at the gadget use list. Just what I was looking for.

Endo999 (talk) 01:37, 5 June 2011 (UTC)

Hi there, I've just taken a look at the User Preferences db page and it hasn't run for 3 weeks. Is this intentional or has something slipped in the db reports?

Endo999 (talk) 22:43, 8 September 2011 (UTC)

[[ISO 3166-2:BQ]] [[ISO 3166-2:CW]] [[ISO 3166-2:SX]]

Hello Dispenser excuse my bad english. This ISO Codes are updatet. http://www.iso.org/iso/iso_3166-1_newsletter_vi-8_split_of_the_dutch_antilles_final-en.pdf. Please can you update your File viewer too? Thanks. --80.142.203.184 (talk) 12:50, 5 June 2011 (UTC)

:BQ, CW, and SX were added to ghel. — Dispenser 04:05, 10 June 2011 (UTC)

::Thanks for this. Good job it works --80.142.157.190 (talk) 20:07, 10 June 2011 (UTC)

That list

Regarding [http://toolserver.org/~magnus/catscan_rewrite.php?templates_yes=Italic_title%0ALowercase_title&doit=1 that list] you told us about during this discussion, I have changed many of those titles by removing the {{tl|Lowercase}} template and inserting {{DISPLAYTITLE:}} in its stead. I didn't get into trouble until I worked on the maNga (album) page. It lead me to the Manga (band) article, which I tried to move to maNga (band). My question to you has to do with the lowercasing of the first letter of these articles. It seems the guidelines tell us to lowercase the first letter only when it is pronounced separately and the second letter is uppercase, as in eBay and iTunes. Many of those articles on the list were more like the maNga articles, and I went ahead and lowercased those first letters in accordance with the WP:COMMONNAME policy, which says to title articles in a way that is familiar to the readers of those articles. So which do you think is correct? maNga or Manga? This is important, because if I've been wrong, then there are a lot of articles that were on that list that I'm going to have to go back and change. – Paine Ellsworth ( CLIMAX )  23:07, 16 June 2011 (UTC)

DAB page for Google phone / Search tool

Many thanks for the DAB page for Google phone, but I wonder if it needs what appears to be a clone of the talk page from the Nexus.

Since you seem to love the challenge of making useful tools, and probably created the above DAB page as a result of seeing the following: This tool idea may interest you. I also discovered that it's possible to set up an article called "Name" without being aware that there is already an existing category called "Name" — the Wikipedia search box doesn't search Category and Sub-Category titles by default (I think it should), and most mere mortals will (a) never work out that it doesn't, and (b) never work out how to use the search box to find if there is a Category or Sub-Category called "Name" ( problem described here ). Categories list many terms with related meanings. It would surely be useful if Wikipedia had a tool to prompt people who attempt to create an article called "Name" that there is already a Category or Sub-Category called "Name", and would they like to view that first (to help them find articles similar to "Name" that already exist in the Category indexes)? LittleBen (talk) 15:16, 19 June 2011 (UTC)

:Typically people create these links by first making a wikilink to a title they think should exists and then redirecting it or creating it. They would've found it if they used the case-insensitive search box. We could add JavaScript to to MediaWiki:Common.js/edit.js to do a search for them when creating a page. But unless your an admins the default action is not to add any new JS.

:Moving on, making a report is probably more useful I could offer. It pretty simple once the collation is set properly:

SELECT GROUP_CONCAT(page_title SEPARATOR "|")

FROM page

JOIN redirect ON rd_from=page_id

WHERE page_namespace=0

GROUP BY CAST(page_title AS CHAR CHARACTER SET utf8) COLLATE utf8_general_ci

HAVING COUNT(DISTINCT rd_title) >= 2

:It takes about 40 minutes to run, create a 2 MB log. — Dispenser 22:03, 20 June 2011 (UTC)

You're invited to the New York Wiknic!

{{Wikipedia:Meetup/NYC/Invite/June 2011}} BrownBot (talk) 19:03, 19 June 2011 (UTC){{clear}}

re: Project statistics

Dab Solver and templates

Hi, Dispenser. In case you hadn't noticed already, a change was recently made to the {{tlx|dn}} template to allow a "date=" parameter; and, to make things even "nicer," a bot is now running around converting all instances of {{tlx|dn}} to read {{Tlx|Disambiguation needed|2=date=June 2011}}. Dab Solver recognizes the old format, but not the new; hopefully it won't be too difficult to adapt to this change. --R'n'B (call me Russ) 14:53, 30 June 2011 (UTC)

: Wow, that makes the regular expression harder AND more language dependent. I've added some relatively untested code to recognizes the most common of the three possible patterns and properly simplify the links. — Dispenser 16:14, 30 June 2011 (UTC)

File viewer

Hello, thank you for this tool : file viewer it's really nice. But I need to know if it currently works only for :fr:template:Cood or also of his French little brother :fr:template:Coordonnées ? Thank you again, Otourly (talk) 14:42, 10 July 2011 (UTC)

:The ghel system is not template based, instead it looks for links to GeoHack. Also, the logs are in an Excel compatible format, which is why a separate tool is needed to view them. — Dispenser 03:07, 11 July 2011 (UTC)

coord-cswiki.log

Hello dispenser, excuse me for my bad english. This site and some other too is shown the errors only like ?curid=886. Is it possible to fix this? --80.142.207.56 (talk) 00:09, 12 July 2011 (UTC)

: I will be looking into optimizing ghel extraction, at which point I will again evaluate the impact of include namespace and title. If you can't wait, you can change your templates to include the &pagename={{PAGENAMEE}} parameter. — Dispenser 05:30, 5 August 2011 (UTC)

Tool

Hello Dispenser, I am translated your documentation from user page for the wiki-pt, this was the last. And I want to know if you can available for the portuguese use your tool? You are planning, for us? Thanks! --Vitor Mazuco Talk! 00:54, 12 July 2011 (UTC)

: My tools should work with all Wikimedia Foundation wikis. I'll look into accommodating differences if posted about it (e.g. different template name). However, there are no plans for translating the Checklinks interface until a major rewrite. There are too many strings and word joining magic to make translating practical. As for Reflinks, translations/i18n existed in the earliest versions and should be added via the pywikibot SVN repository. — Dispenser 22:05, 12 July 2011 (UTC)

ghel LOAD DATA INFILE error

Hi Dispenser, the File Viewer has a mistake. See the bottom of all country-pages. --Knochen (talk) 17:13, 22 July 2011 (UTC)

:Their was an conflict with mysqldump and the new "local-infile" option yesterday. The supposed fix (prefixing "loose-") which broke it. It is finally fixed by explicitly setting local_infile=True in the connection handling part of the python code. The upside to all this is I can re-load the data files. — Dispenser 20:14, 22 July 2011 (UTC)

Grammar error in automated peer reviewer

Hi Dispenser, in the text about redundancy in the automated peer reviewer, "Vague terms of size often are unnecessary and redundant" should be changed to "Vague terms of size are often unnecessary and redundant". Thanks. Graham87 03:00, 26 July 2011 (UTC)

: My tool is only a loader, so you'll need to change the original source at User:AndyZ/peerreviewer.js (line 603). — Dispenser 15:33, 3 August 2011 (UTC)

click tracking

Thanks for pinging me on the click-tracking GLAM stats idea that you put on the village pump. I've forwarded a link to that to the cultural-partners mailing list if anyone technical there wants to get involved with the idea. I can say however that Stats/metrics/analytics things are a big and hairy piece of software... I've been told that, whilst the WMF is indeed building up it's capacity in this field, GLAM-specific stuff is not first priority. I'm working on alternatives to this but I can say it's not going to happen in the short term. That said, it doesn't mean your proposal necessarily has to go through centralised development things like the WMF, so don't let that stop you if you've got consensus and someone to code the thing :-) Wittylama 06:06, 26 July 2011 (UTC)

What File viewer do

Hello,

I would like to have more information about your greatfull tool. For the error region invalid, the tool detect an error based on ISO_3166-1 or ISO_3166-2 ? I know we could put the ISO_3166-2 code but I don't know that if my coordinates are out, the tool detect the subdivision problem or only for countries ? What about the case of international lakes or mountain, could we write for exemple for the ":fr:Mont-Blanc" FR-74/IT-23 ? How will react your tool ? And what about three-regions (or more) objects ? Regards, Otourly (talk) 19:07, 26 July 2011 (UTC)

:Could you also do a log for frwiki like this one which is for dewiki : iwcoord Thanks a lot :) Otourly (talk) 20:42, 27 July 2011 (UTC)

::Did you understand my questions ? Otourly (talk) 16:02, 2 August 2011 (UTC)

: We use ISO 3166-2, but the question is more about implementation support. The current implementation limits region: field to 60 bytes which is enough for 10 region codes with subdivisions separated by slashes. However, aside from the abandoned regioncheck script that checks the point is inside the region, little software uses this information. The original intent was to give GeoHack the ability to hide services that did not have coverage, but this has become a maintenance nightmare (dewiki didn't notice half the links were duplicates). Due to its relative uselessness and that an automatic region locator is working, I am considering deprecating the field.

: I plan to rewrite the interwiki coordinates (iwcoord) script as tool. This will be far faster (enwiki currently takes 2.3 hours!), perform some analysis and suggest text to copy.

: My free time is limited and divided among 4 major tools and many various smaller projects. My current focus is on introducing new editors through disambiguation (dab solver). If you're looking for a quicker response then I suggest that you contact me on IRC. As a rule of thumb, it is usually quicker for me to write code then to reply to people. — Dispenser 18:42, 3 August 2011 (UTC)

::OK no problems :) Thank you :) Otourly (talk) 18:13, 4 August 2011 (UTC)

:::It's not really a problem but File viewer make an error for the "region:SS" that's correspond to South Soudan. Otourly (talk) 17:04, 25 August 2011 (UTC)

:::: Added. Really? The ISO can't go two months without adding another code. — Dispenser 14:36, 26 August 2011 (UTC)

stable.toolserver.org will be ignored soon

Hello Dispenser, the File Viewer shows this error on all sides. Can you fix it? --Knochen (talk) 16:03, 12 August 2011 (UTC)

:Well not on all site, some languages are much more progressive :-). Anyway, it's commented out now as got what I need (who's still using stable.toolserver.org). — Dispenser 21:37, 13 August 2011 (UTC)

::I'm interested in a list for dewiki. Can you search for "http://stable.toolserver.org/geohack" again when running your script tomorrow? Thanks. --89.204.152.53 (talk) 15:48, 14 August 2011 (UTC)

:::I'm likely not going to re-enable the legacy URLs due to performance. You can use w:de:Special:LinkSearch/http://stable.toolserver.org/geohack/ (and w:de:Special:LinkSearch/http://toolserver.org/~magnus/geo/geohack.php) instead. You may also want to check out WT:WikiProject Geographical coordinates/to do. — Dispenser 19:23, 14 August 2011 (UTC)

::::I've checked Special:LinkSearch before. There are listed 70 pages but I remember your logfile showing more than 700 some days ago! That is why I would like to check those again. --89.204.137.106 (talk) 19:56, 14 August 2011 (UTC)

:::::Temporarily re-enabled, as I would like to see where the flaw is. — Dispenser 20:34, 14 August 2011 (UTC)

::::::Thanks. --89.204.137.106 (talk) 20:39, 14 August 2011 (UTC)

::::::: Results are in: 60 for dewiki and 379 for enwiki. So LinkSearch is accurate and I'll disable it once more. — Dispenser 14:08, 15 August 2011 (UTC)

::::::::So there must have been some change at dewiki in the last few days. Thanks again --89.204.152.53 (talk) 14:24, 15 August 2011 (UTC)

PDFBot edit summaries

Could you please take a look at Wikipedia:Bot owners' noticeboard#General notice to bot owners about edit summaries and see if the suggestions might apply to your bot? Feel free to add your own suggestions and comments there too. Headbomb {talk / contribs / physics / books} 21:12, 21 August 2011 (UTC)

: PDFbot had already made progress by counting the number of PDFs worked and the three other states to spit out one of the six unique edit summaries. However, I am more keen on spending the time improving my web tool to better summaries, since the operators are indifferent of improving the default. — Dispenser 21:09, 26 August 2011 (UTC)

Strange edits left by Dabfix

Hi. I was using Dabfix and it left behind some rather strange syntax relating to the intro. Another editor has fixed most of these entries (like [http://en.wikipedia.org/w/index.php?title=John_Cunliffe&curid=21797040&diff=447658310&oldid=447599898 so]). Just a glitch? Tassedethe (talk) 15:14, 31 August 2011 (UTC)

:Code was added to {{refer}} recently without check that it subst: correctly. This has been now fixed. — Dispenser 17:49, 31 August 2011 (UTC)

Access granted

class="wikitable sortable" style="float:right;"

|+ September Dab Challenge Statistics

! User

! Fixes

! EditHours

! Edit/min

! Bonus

R'n'B

| 3801

| 58.7

| 1.08

| 17%

Woohookitty

| 2649

| 28

| 1.58

| 1%

JustAGal

| 2225

| 121

| 0.31

| 99%

Squids and Chips

| 1669

| 53.5

| 0.52

| 24%

JaGa

| 898

| 26.3

| 0.57

| 20%

J04n

| 730

| 45.2

| 0.27

| 64%

William Avery

| 657

| 29.2

| 0.38

| 4%

StAnselm

| 467

| 30.7

| 0.25

| 38%

Bte99

| 425

| 31.7

| 0.22

| 83%

Gongshow

| 407

| 6.3

| 1.07

| 1%

JamesAM

| 328

| 31.2

| 0.18

| 40%

C777

| 257

| 7

| 0.61

| 0%

X201

| 226

| 18.8

| 0.2

| 85%

69.248.62.131

| 196

| 21.3

| 0.15

| 93%

TimBentley

| 190

| 6

| 0.53

| 15%

BD2412

| 181

| 7.3

| 0.41

| 27%

Odie5533

| 181

| 7

| 0.43

| 44%

GoingBatty

| 154

| 6.7

| 0.39

| 19%

Koumz

| 129

| 7

| 0.31

| 31%

Logan

| 117

| 12.7

| 0.15

| 93%

Hey, I just wanted to let you know all of my Toolserver data has been ported over to u_jason_p, so everything is now available for you to play with. Let me know if you have any questions. Cheers, --JaGatalk 06:22, 1 September 2011 (UTC)

/* Count Dab Challenge by WikiProject */

SELECT pb_title, COUNT(*)

FROM u_jason_p.all_dab_links

LEFT JOIN u_dispenser_p.projectbanner ON pb_page=article_id

JOIN u_jason_p.contest_dabs ON c_id=dab_id

GROUP BY pb_title

ORDER BY COUNT(*);

: Now we can make personalized Dab Challenge lists. This was previously the number one question: "Why haven't I gotten any points?" Now I could only ping something when a users saves so we'd have a live leaderboard. — Dispenser 21:09, 1 September 2011 (UTC)

::Cool. I was wondering what you and Russ could do with the data, once I finally put it out there. I'm looking forward to developments. Cheers, --JaGatalk 22:25, 1 September 2011 (UTC)

::: I've always wonder how much time is spent disambiguating. It seems people can be competitive even if they only have weekends. — Dispenser 18:48, 4 September 2011 (UTC)

::::Ha! Looks like I'm not getting enough bang for my buck. Somehow I'd assumed Woohookitty got closer to a 40-hour work week in. BTW, I really like the progress bar in Dab Solver. --JaGatalk 07:33, 5 September 2011 (UTC)

::::: Minor table update. I've come up more columns, but leaving them out too keep it consistent. — Dispenser 12:43, 12 September 2011 (UTC)

:::::: Falling behind updating this table. Anyway, do you keep the rev_id + fix count for previous months? — Dispenser 07:02, 21 September 2011 (UTC)

:::::::No, I don't archive any thing but the top four contest results. I've often thought it would be nice to have a grand running total for each person. --JaGatalk 16:36, 21 September 2011 (UTC)

:::::::: Looking back: I think its important to have something where people can feel a sense of accomplishment even if only a few hours are spent. The leader board seemed like a bit of deterrent towards that. Achievement may be the way to go, although coming up with them is a bit limiting due to decreasing nature of the system.

:::::::: And now stats! The Video Games WikiProject is the first Project to finish their list. JustAGal traveled to the most dab page (1099) with R'n'B and J04n in a distant second at 490 and 454 dab pages. While not as impressive Logan did manage to fix 117 dab links without ever repeating one. — Dispenser 04:24, 4 October 2011 (UTC)

So many questions

I've got several questions/comments that have been building up lately:

  1. Bummer about the loss of the redlink listing in Dab Solver. That was an excellent feature. I do use the redlink prefix tool you provided, but it just isn't the same. Any idea why the queries slowed down? I notice you've placed a time limit on the redlink queries.
  2. In the status bar for Dab Solver, you use the phrase "You're rival is" but it should be "Your rival is".
  3. Headbomb mentioned a "Disambiguation bot" that places {{tl|dn}} inline templates in articles and does something with talk pages. Is this WildBot? He was asking if I were interested in running it since I have a TS account. I'm willing to give it a shot, but I know little about it.

Thanks, --JaGatalk 05:15, 13 September 2011 (UTC)

:# I suspect a back end change is to blame. Mostly it is related to the CAST(pl_title AS CHAR CHARACTER SET utf8) that isn't present in redlink.py. The time limit was there before, but I've had to a lot less generous so it doesn't take 2 minutes to load per dablink. As it stands, the medium generation for the related map+redlinks suggestion has increased from 5 to 11 seconds even with the 20 timeout.

:# Corrected it. Still need to make it more compelling and nicer to look at, but first landing page needs an auto-updater.

:# You'll find it at /home/dispenser/public_html/cgi-bin/i_am_a_bot.py, and its runnable from the web interface. It just something I quickly hacked up to do regex link matching.

: You can contact me on IRC (see top) if you have questions. — Dispenser 19:49, 13 September 2011 (UTC)

/* Estimate article transclusion count for templates (and other embedding) */

SUM((SELECT COUNT(*) FROM templatelinks

JOIN page AS trans ON trans.page_id=tl_from + 0 AND trans.page_namespace=0

WHERE tl_namespace=refer.page_namespace AND tl_title=refer.page_title)) AS trans_count

:: Apparently the query optimizer suddenly changed its mind. Before it looked up the transclusion then discarded non-mainspace results. Now it looks up which of the two million articles have templates and try matching those against the results in the query. Since we can't use USE/FORCE INDEX, I came up with a little hack of tl_from + 0 to prevent the optimization from happening. Now its back to 0.8 seconds from 5 hours. — Dispenser 02:33, 14 September 2011 (UTC)

Dabsolver error

The following is now appearing just above the #workarea element on each dabsolver page:

UnboundLocalError("local variable 'rival' referenced before assignment",)

--R'n'B (call me Russ) 21:28, 16 September 2011 (UTC)

: Fixed (only now replying). Too focused on optimizing the newbie case that I overlooked the obvious case on the other end. Probably should add easter eggs... — Dispenser 05:12, 22 September 2011 (UTC)

A barnstar for you!

style="background-color: #fdffe7; border: 1px solid #fceb92;"

|rowspan="2" style="vertical-align: middle; padding: 5px;" | 100px

|style="font-size: x-large; padding: 3px 3px 0 3px; height: 1.5em;" | The Technical Barnstar

style="vertical-align: middle; padding: 3px;" | I, Sarah, hereby award Dispenser this barnstar for all the amazing tools, programs, ideas, and smartness that Dispenser..well...dispenses. Thank you for all you do for Wikipedia and it's related projects!! SarahStierch (talk) 02:49, 22 September 2011 (UTC)

Talkback

{{talkback|Wikipedia talk:Signatures|SQL report of accounts violating guidelines|ts=08:01, 23 September 2011 (UTC)}}

Your further input has been requested. Thanks. Kudpung กุดผึ้ง (talk) 08:01, 23 September 2011 (UTC)

New dabsolver.py issues

Since about 1 Oct, clicking on "Unlink" is converting Link to Link instead of unlinking it; and clicking on "I don't know / Tag {{dn}}" is not doing anything at all. --R'n'B (call me Russ) 20:23, 2 October 2011 (UTC)

: Fixed. The issue was a switch to protocol relative URLs tripped the interwiki link maker which is stricter about proper page titles. This is the reason why it doesn't come up in testing earlier. — Dispenser 04:20, 3 October 2011 (UTC)

Hello there!

I've tried Dap Solver in En.Wiki, it's amazing, I'd like to translate it into Arabic.Wiki! How about that??

Thanks :) Sean (Hit) 06:39, 3 October 2011 (UTC)

:Just create a new section on /Dab solver, copy the code from the Translating to danish section (not the most intuitive title), and translate the blue parts. I'll try to handle the rest, although I expect to run into trouble since this is my first rtl conversion. Some notes: rtl is not well supported by the APIs (i.e. no layout flipping), the landing page is not translated due to the amount of work, and the disambiguation contest run by User:JaGa is English only and likely stay that way due to resource limitations. — Dispenser 06:24, 4 October 2011 (UTC)

No “bodytext” crashes rdcheck

Hi, just letting you know your rdcheck tool seems to be crashing on the strcat page: tools:~dispenser/cgi-bin/rdcheck.py/strcat. It gives a “substring not found” ValueError on line 90; I think it is failing to find a <!-- bodytext --> comment tag in the HTML page. Perhaps Wikipedia has recently changed the way the pages are coded, because I recently edited that page. Vadmium (talk, contribs) 01:20, 5 October 2011 (UTC).

Sure enough, I checked your talk page before adding that post and it was fine; now rdcheck crashes on your talk page as well :). I wonder if you could get away with searching the whole page for element identifiers without limiting to the “bodytext”. Vadmium (talk, contribs) 01:23, 5 October 2011 (UTC).

IE error on Dab Solver

Hey, I got this message a bit ago:

:Well Ja Ga: Fixing my new article "Sergio Franchi discography" using your Bot seemed simple enough. I Went through the fix, but get a message that Internet Explorer is preventing cross-site scripting. Have you a solution, or should I just remove the Internal link? Thanks, CatherineCathlec (talk)

Do you know how to help her, or what went wrong? Thanks, --JaGatalk 02:30, 28 October 2011 (UTC)

:She's referring to the XSS Filter notice (users tend to miss it popping up). Her options are to disable the XSS Filter or use a different browser. Bug 32013 addresses with fixing this. — Dispenser 03:02, 10 January 2012 (UTC)

Regions for coordinates templates on Commons

Hi Dispenser. I have opened a request for bot approval for adding region info to coordinates templates on Commons. I think it is useful metadata, including for your coordinates databases (coord_commonswiki_p). Can you give your opinion there please? Regards. emijrp (talk) 18:35, 28 October 2011 (UTC)

Dab Solver error

I am getting this error when I follow the Computers link on the tool,

A problem occurred in a Python script. Here is the sequence of function calls leading up to the error, in the order they occurred.

/home/dispenser/public_html/cgi-bin/godab.py in ()

18 if './' not in topic:

19 try:

=> 20 f = open(path, 'r+')

21 size = os.path.getsize(path)

22 #offset = int(random.uniform(0, size - min(4096, size/2)))

f undefined, builtin open = , path = '../temp/dabs/challenge/WikiProject_Computing'

: [Errno 90] Number of symbolic links encountered during path name traversal exceeds MAXSYMLINKS: '../temp/dabs/challenge/WikiProject_Computing'

args = (90, 'Number of symbolic links encountered during path name traversal exceeds MAXSYMLINKS')

errno = 90

filename = '../temp/dabs/challenge/WikiProject_Computing'

message = ''

strerror = 'Number of symbolic links encountered during path name traversal exceeds MAXSYMLINKS'

/home/dispenser/public_html/cgi-bin/tracebacks/tmpMxaALI.html contains the description of this error.

Mo ainm~Talk 21:12, 29 October 2011 (UTC)

:Fixed. It was ln -s ./ challenge not ln -s challenge ./. — Dispenser 21:17, 29 October 2011 (UTC)

:: So you did, that was quick, good work. Mo ainm~Talk 21:21, 29 October 2011 (UTC)

dab_solver.py grammar

"Your closing in on" should be "You're [or You are] closing in on".... --R'n'B (call me Russ) 15:37, 1 November 2011 (UTC)

  • And, "[Someone]‎ is now -218 behind of you" should be "[Someone] is now 218 behind you." --R'n'B (call me Russ) 14:00, 3 November 2011 (UTC)

ghel gc_primary field

Hi,

I really like the ghel tool and its regular dumps that are available via your tool server account. i have a question concerning the ghel tool: i do not completely understand the field gc_primary within the output table. initially i thought that this field indicates whether there is a link on top of the wikipedia page which opens the openstreetmap map. but there are pages within the german wikipedia which are set to gc_primary = 1, but do not have such a link. e.g. de.wikipedia.org/wiki/CDU (page_id = 3956889 = gc_from). so what is the actual supposed meaning of gc_primary?

Regards,

Stefan — Preceding unsigned comment added by 141.30.4.138 (talk) 17:01, 1 November 2011 (UTC)

Dispenser (via IRC): The gc_primary is meant to indicate the desiganted coordinate for the topic. Unfortunately, the syntax (absents of &title=) is the same for when people forget to name referenced locations in the article. I don't understand why, but some people despise title coordinates.

A barnstar for you!

style="background-color: #fdffe7; border: 1px solid #fceb92;"

|rowspan="2" style="vertical-align: middle; padding: 5px;" | 100px

|style="font-size: x-large; padding: 3px 3px 0 3px; height: 1.5em;" | The Original Barnstar

style="vertical-align: middle; padding: 3px;" | Thank you for showing me how to use CatScan :) --Sp33dyphil ©© 06:42, 10 November 2011 (UTC)

dewiki-CoordinateNO

Hi Dispenser, for this page (dewiki-CoordinateNO.log) there is no update - last at 14 Sep 11. Can you fix it? --217.246.216.131 (talk) 02:32, 12 November 2011 (UTC)

: Fixed. It was the result of a broken symlink. — Dispenser 21:05, 12 November 2011 (UTC)

Red Links and Dab solver

Howdy. I've created a database u_tb_dab_p on s1, in which you'll find a single table rls ( redlink varbinary(255), suggestion varbinary(255) ). I've populated it with an initial million rows for you to be going on with. This reduces our interface to:

SELECT suggestion FROM u_tb_dab_p.rls WHERE redlink = 'some red link';

An index has been added to support the above query. You're welcome to amend the table structure and contents to suit your needs. I'll plan on maintaining it roughly every 6 months. We can tinker with the quantity vs quality of suggestions once we see how it works. - TB (talk) 07:52, 28 November 2011 (UTC)

dabfix.py stalls on BMR, POD and GM

I'm not sure if it's because they're short article names or because they're uppercase, but dabfix.py never returns HTML to edit these dabs. Josh Parris 00:16, 1 December 2011 (UTC)

I finally got a return [on BMR], and when I click on "Show details" I get:

ERROR: Redlink FullText search: ProgrammingError(1146, "Table 'u_dispenser_p.redlinks_enwiki' doesn't exist")

but the full HTML page wasn't returned. Josh Parris 12:37, 26 December 2011 (UTC)

: Fixed Turned out that a regular expression in getsummary() was too greedy and accidentally spanned multiple lines. — Dispenser 19:30, 9 January 2012 (UTC)

Dab Solver question

Any idea what happened {{diff|Captain_Bill_Bellamy_MC|463655582|463537420|here}}? Brackets were replaced by hashes for many wikilinks. The problem seems to have gone away since, but an editor was asking me about it (via DPL bot's {{tl|dablinks}} templates). --JaGatalk 15:50, 2 December 2011 (UTC)

:Haven't noticed my canvasing for Bug 32013 for the last month? What's happening: When a IE8/9 user follows a link or submit a form to a different domain the regular expression based "XSS Filter" attempts to mangle and disable JavaScript in areas matching the GET/POST data. It is sneaky as mangling is done after the page has been served and only affects the edit box leaving the preview or diff view intact. WMF can disable it by sending X-XSS-Protection: 0 header or users can disable it following the instructions linked from my tools. — Dispenser 22:25, 3 December 2011 (UTC)

::Actually, I did notice the messages but forgot to check it out. (I went and voted for it BTW - or at least added myself to the CC list, which I assume is the same thing.) Thanks for the explanation. --JaGatalk 00:20, 4 December 2011 (UTC)

Protocol relative Link to GeoHack

Hello Dispenser, how did your tool "LocateCoord" find the coordinates? It is a problem for your tool, when the weblinks to the GeoHack are protocol relative or HTTPS only? When yes, it is possible to fix your tool? Thank you. Der Umherirrende 18:55, 5 December 2011 (UTC)

: I have improved GeoHack's HTTPS support and switch to protocol relative URL on :de:Template:GeoTemplate where possible. As for the ghel parser and database (locateCoord.py is only a front-end), the revised code using el_index instead of el_to.† This query could run twice as slow, but reproducing it consistently is difficult as the regular query run varies between 2-24 minutes for dewiki. Decreased performance is unacceptable with the script already running for three hours daily.

: † Protocol relative URLs are stored twice in MW database with a "http:" and "https:" normalized search field (el_index). — Dispenser 07:52, 7 December 2011 (UTC)

::Thanks for your changes. Maybe a DISTINCT can help here, to avoid duplicate entries in the result set. Der Umherirrende 20:38, 7 December 2011 (UTC)

:::Using DISTINCT entails reading the data twice and discarding half which is at best O(n×log n). The row look up for el_from should only be O(log n) and be faster, bar unexpected caching. So I'll be testing the el_index method for all wikis over the next week. — Dispenser 07:47, 10 December 2011 (UTC)

::::Looking at the time data from the logs show no correlation to the introduction el_index search method. I've also managed to save 60 minutes of CPU by writing an incremental updater for commons. So we should be good for the switch to HTTPS on my end. — Dispenser 23:47, 7 January 2012 (UTC)

Yet another dab_solver.py question

Hi. How come dab_solver.py always tells me "BD2412‎ is now XXX behind you"? It is always BD2412, no matter how far behind me he is or how many other users are between the two of us. Right now, PhnomPencil is the user just behind me, but I'm still getting BD2412 messages. (Maybe it's psychoanalyzing me....) --R'n'B (call me Russ) 15:52, 7 December 2011 (UTC)

: It's a pre-generated list to match active users with a buddy. The algorithm is similar to WikiStalker as it attempts to pair you with someone that you interact with recently (30 days/recentchanges table). I last updated the list November 2 and since I was still playing around with the implementation when my laptop power connector broke I never got around to automating it. — Dispenser 20:29, 7 January 2012 (UTC)

[[Fawlty Towers]]

Hi. Pending any further discussion on the appropriateness of your bot calls in this article, I've reverted 'em because they appear unnecessary and inappropriate. The (now somewhat archaic) word 'cretin' is adequately defined in Wikipedia. I would say that meaning No. 2 is actually the more prevalent since the medical substitution of specific terms such as Congenital hypothyroidism. Anyway, the provided Wikilink for Cretinism perfectly disambiguates the former (discontinued) medical usage and the ironic pejorative usage (which is itself, of course, subjected to implied irony by being placed in the mouth of Basil Fawlty). Also, the term 'waiter' is fully disambiguated by the provided Wikilink. It seems to me that the use of bots for such purposes has significant limitations when dealing with subjects involving habitual English usage and important irony, and that you should enter a proper discussion before making such calls. But maybe you can persuade me otherwise. Cheers,01:13, 8 December 2011 (UTC)

: Someone's already {{diff|Fawlty_Towers|464821341|464673304|correctly disambiguated the links}} for you and I've added the Wiktionary (it's a sister project) to the disambiguation page. The reason why I added the dab buttons is because the page still has a WildBot notice (albeit for a different link) and has over 10 active watchers (40 in the last 3 days). — Dispenser 03:44, 10 December 2011 (UTC)

Hah!

Nice userpage :D ResMar 03:06, 21 December 2011 (UTC)