User talk:Duesentrieb/CheckUsage

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

This page is for discussion about CheckUsage [1], a tool written by User:Duesentrieb. If I do not respond, please leave a quick note on my talk page at the german wikipedia.

New CheckUsage[edit]

moved here from commons:User talk:Duesentrieb -- Duesentrieb 17:01, 9 January 2006 (UTC)

Does your new version of CheckUsage at actually check all WikiMedia wikis? If so, that solves one big problem with deleting redundant pics. Is there any reason not to switch commons:Template:Redundant to link to your version? Thanks! commons:User:dbenbenn 17:18, 26 December 2005 (UTC)

It checks all projects except a few internal ones (like the fundraising wiki, wikimania-wiki, etc), if you tell it to check everything.
I don't know of any reason not to switch, except that I don't have time to respond to complains right now... there may still be problems no one has noticed yet. So, I'd suggest to wait a few more days with changing the template. -- Duesentrieb(?!) 00:28, 27 December 2005 (UTC)
One (slight) problem: links from "Wikipedia:" pages on EN appear as "Project:". That's not a huge problem, since the link works either way. But links from "Portal:" pages on EN appear as "100:", and the link is broken. See, for example, [2]. commons:User:dbenbenn 18:59, 1 January 2006 (UTC)
"Project:" is the canonical namespace name, which works on all wikis. The problem with Portal, etc, is that custom namespaces are different in all projects, and they are not stored in the database. I have no way of knowing them, i fear. -- Duesentrieb(?!) 01:59, 2 January 2006 (UTC)
Too bad MediaWiki doesn't include a default namespace name for custom namespaces. If you can't figure out the names, you certainly can at least handle the case better. A link to 100:Aviation/Selected picture/October is not useful. How about unlinking in this case, and adding a disclaimer, something like
100:Aviation/Selected picture/October (to find this page, go to the image description page)
Dbenbenn 18:19, 4 February 2006 (UTC)
Currently, i'm thinking about extracting the extra namespaces from all projects using Special:Export, and putting them into the toolserver database, linked to the wiki table. That would be the cleanest solution, I think. Check again in a few weeks... -- Duesentrieb 14:26, 5 February 2006 (UTC)


I will be thankful, if you explain, if it is OK, if I will use a pic, used in English Wikipedia (Turkish Van), in Russian Wikipedia in Russian article about Turkish Van, because this pic was made in 1950-ies, and gives the imformation about the phenotype of the first registered Van cat.--Zara-arush 10:59, 13 July 2009 (UTC)

Ideas for improvement[edit]

I find some features of the old Check-Usage from Avatar quite good. So I think the interface could be improved the following:

  • The (optional) posibility to check more than one image at once (bulk-check).
  • Results in pasteable wikisyntax ([[:de:Image:foobar.jpg]]).
  • Results are expanded by default up to a certain limit of results: They get not expanded for a single wiki if there are more than 20 hits within one wiki and they get not expanded globally if the results of all expanded wikis with up to 20 hits are more than 200 (or 100) hits. And possibly an "expand all" button (switch that changes to "collapse all" if everything is expanded).

That's all I think. Arnomane 18:08, 27 January 2006 (UTC)

  • Button not to check discussion pages. For deletion sysops they are useless. --Paddy 23:44, 10 February 2006 (UTC)
  • Ability to remove the ?uselang=en (is sometimes annoying because visited marking) (and some people would maybe like another language) Could maybe be a cookie, and if this is empty, no ?uselang is appended. -- 10:10, 17 September 2006 (UTC)

check-usage image tab fix?[edit]

Wow, where did this come from? A true CheckUsage?! My dream!!

I didn't even notice, because I use the check-usage tab that I installed. Now how can I fix it so it really checks all (ie points to Duesentrieb's tool instead of Avatar's)? I can't remember where I got it from, but my code is at commons:User:Pfctdayelise/monobook.js. Thanks, pfctdayelise 01:35, 15 February 2006 (UTC)

bug in ja:?[edit]

Looking at [3], it lists two pages in ja:. But I know this image is not in these pages because I commented them out yesterday. The image is Image:Björk performing at the Olympics in Athens.jpg and you can see for yourself it's not there. I commented the image out of many projects, and it works fine for all the rest, even he: (ie, they don't show up in CheckUsage). So I wonder if this is just a problem in ja:? Interestingly, they correctly don't show up in Avatar's old CheckUsage: [4]. pfctdayelise 02:08, 15 February 2006 (UTC)

apperently we are not getting any updates from the asian cluster to the toolserver. This is not easy to fix... I hope we'll manage it soo, though. -- Duesentrieb 14:17, 14 March 2006 (UTC)[edit]


when scanning with the function "all wiki's", the tool doesn't scan . Could you add it on the list? (and maybe also or fr. etc? ) Effeietsanders 19:18, 13 March 2006 (UTC)

last I checked, nl.wikimedia was a "closed" wiki, and I did not have access to the database of that wiki. Maybe I can convince people that I can get access to everything but page text, not sure. -- Duesentrieb 14:18, 14 March 2006 (UTC)
It is open for everyone with an useraccount. It's indeed not editeble for anonymous users. But everyone can create useraccounts. (but everybody can view it afaik, also anonymous) Effeietsanders 14:24, 14 March 2006 (UTC)
Well, to me the relevant question is if I can read the DB on the toolserver. I'll try to get Kate to open it up, but don't hold your breath... -- Duesentrieb 11:47, 15 March 2006 (UTC)
It's fixed, say thanks to Kate :) -- Duesentrieb 01:22, 16 March 2006 (UTC)
Thank you a lot as well! :) Effeietsanders 12:52, 17 March 2006 (UTC)

not recognising local picture is different[edit]

Usually this is not a problem, but this was weird:

[5] doesn't have the usual "different image" message. pfctdayelise 09:24, 24 March 2006 (UTC)

That's because it doesn't find the image on the commons at all - it was uploaded today, and the replication lag of the toolserver is currently > 20 hours (I guess replication broke during the outage yesterday). -- Duesentrieb 18:15, 24 March 2006 (UTC)
Oh, right... duh! I forgot to check that. pfctdayelise 00:26, 25 March 2006 (UTC)

Bulk checking[edit]

A couple of ideas for the future:

  • It would be good to have a direct link from CatScan to the Bulk checker, so you can automatically check everything in a category (super useful for deletion categories). I didn't realise that the Bulk checker could handle input like ' Images with unknown source as of 2 March 2006 Image:ARADO01.jpg ' with all the extra rubbish, very handy. I think before, I was manually removing the rubbish.
  • Along the same lines, the Bulk checker output would be vastly more useful organised by Image rather than Project. Then you can see immediately if any images are not being used at all, they can be deleted straight away. (Do the easy cases first. ;))
  • While I'm here... this output doesn't seem that useful, I mean listing the base category. It seems kind of obvious and just takes up space.

Thanks :) pfctdayelise 14:57, 31 March 2006 (UTC)

Thanks for your comments.
ad 1: uh, it does work with all the rubbish? How strange :) Perhaps a direct link would be useful - but i'm not too keen to make heavy queries too easy...
ad 2: i can only check one project at a time. organizing by image would mean that I can produce output only after everything has been checked... and it may also take up a lot of memory. Not sure if i'll do it.
ad 3: It's useles only because there are no subcategories. Perhaps I could make a special case for that. But doing that seems inconsistent. -- Duesentrieb 23:25, 2 April 2006 (UTC)
ad 2: that's a shame. But if we can get a delinker bot on commons, it won't be necessary, so I think that's a better solution.
ad 3: But see [6]. It seems to me that it's always useless information, because you always already know that whatever files show up are in the base category. You have to supply a base category - there's no other way to make it work. Well, that's my opinion anyway. :) pfctdayelise 13:41, 3 April 2006 (UTC)
No, they could be in a subcategory of the base category. To know in which one would be useful. consider [7] -- Duesentrieb 00:21, 7 April 2006 (UTC)

Workaround für die en. - / Asien-Datenbank-, äh, "Situation"...[edit]

Hi Daniel,

Wie wir ja alle wissen, sind die Toolserver-Anwendungen nicht mehr so ganz so der Wahnsinn, wie se es mal waren ;) Ich hoffe, das wird demnächst behoben, schliesslich sind die Tools ja das beste Ding seit MediaWiki. Okay, zur Sache: Wie wäre es, wenn du, solange das automatische Parsen nicht geht, einfach eine Checkbox einbaust ("Include unreachable databases" o.ä.; ich würde das sogar standardmässig aktivieren) und dann, falls gewünscht, einfach unten am Ende der Liste statische Links zur jeweiligen Beschreibungsseite in der Form " unknown", " unknown" und so weiter anbietest - da kann man dann draufklicken und selber nachsehen, und das wäre immerhin besser als nichts. Im Moment mache ich das per Hand, das ist aber mühselig. --Grm wnr 15:48, 25 April 2006 (UTC)

Hm, ja, wenn das noch länger so bleibt, könnte ich das einbauen. DaB hatte haber letzte Woche schon mit der en_wp getestet, und das sah schon ganz gut aus... Vielleicht wird's ja jetzt was, muss mal gucken, das der aktuelle Stand ist - ich war ein paar Tage unterwegs. -- Duesentrieb 21:24, 26 April 2006 (UTC)


There seems to be a bug when the article that contains the image is a portal?! I tried [8] and the link to the "Portal Diskussion:Augsburg" does not work because there is a "+" instead of a "_". -- 21:41, 7 May 2006 (UTC) (German wikipedia: User:Thgoiter)

Ja, scheint ein Bug zu sein... hm, ist ein bisschen komisch... Muss ich mir nochmal angucken. -- Duesentrieb 23:31, 7 May 2006 (UTC)
Der Bug besteht nach wie vor. Liegt am en:url encoding, weil "+" dabei ein reserviertes Zeichen ist, und somit korrekt in ein Underscore umgewandelt wird). Das CheckUsage-Tool müsste also die reservierten Zeichen in Bildnamen zwischendurch in Hex-Werte umwandeln. Workaround: Händische Umwandlung in Hex-Zeichen.
Bsp. für Tool-Eingabefeld "Image to check":
  • "AUT_mail_advice_of_delivery_(front+back).jpg" funktioniert nicht
  • "AUT_mail_advice_of_delivery_(front%2bback).jpg" funktioniert schon
Den Tool-src fand ich übrigens unter downloads nicht, aber es ist auch so klar, was das Problem ist. mfg de:Contributor / 15:40, 20 September 2006 (UTC)

Turn the asian ones back on[edit]

The asian wikipedia's are turned off, even though they now work. You probably want to fix this. 02:52, 13 May 2006 (UTC) (actually: w:en:User:JesseW/not logged in)

Yes, please :) pfctdayelise 04:44, 13 May 2006 (UTC)
yea, i did that a few hours ago - i though i did yesterday, but apperently i got sidetracked ;) -- Duesentrieb 18:22, 13 May 2006 (UTC)


What about showing a thumb of the checked image on the results page? That would be nice. -- 20:00, 14 May 2006 (UTC)

will do, thanks for the suggestion -- Duesentrieb 13:24, 15 May 2006 (UTC)


It seems to be b0rking at 79% with boardwiki pretty bad: "Error in fetchObject(): Table 'boardwiki_p.image' doesn't exist". I never noticed it even checked boardwiki before... pfctdayelise 08:57, 18 June 2006 (UTC)

strange... apperently, that database was deleted from the toolserver or somethign went wrong with permissions. I have makred it as broken for now. -- Duesentrieb 12:30, 18 June 2006 (UTC)

Sega master system II

Misfeature? Truncation of long filenames[edit]

The link from the check usage tab (on a Commons image) took me to:

which searched a truncated filename, and incorrectly showed no usage:

697 wikis searched. Bristol_published_by_A._Fullarton_ is not used anywhere!

Pasting the full filename directly into the form took me to:

which correctly listed usage of the file:

21 wikis searched. Bristol_published_by_A._Fullarton_&_Co_in_1866_edited.jpg is used on 2 pages in 2 projects.

Maybe there are some changes needed to code at Commons.

--commons:User:InfantGorilla 11:00, 10 October 2006 (UTC)

  • I have observed the same problem with many filenames with ampersand (i.e. "&" sign) inside. Julo 10:44, 28 October 2006 (UTC)

This appears to be a problem with the way to "extra tabs" on commons work - the tool itself handles "&" correctly in filenames (however, it seems to have a problem with "+", see below). Please ask the person who wrote the tab integration (Arnomane, iirc). Btw: I don't see the "extra tabs" at all any more. Strange. -- Duesentrieb 13:05, 2 November 2006 (UTC)

"plus" sign in the filename[edit]

If I want to check Image:Adam_Michnik_+_Tadeusz_Mazowiecki.JPG file, the tool works with "Image:Adam_Michnik___Tadeusz_Mazowiecki.JPG" name - with plus sign replaced by underline (as result, there are three underlines close by).

Regards Julo 10:22, 28 October 2006 (UTC)


Strange - this is somehow related to the fact that "+" is used to encode spaces in URLs. This seems to be handeled incorrectly by the "extra tabs" JS code on commons. However, when pasting the filename directly, "_+_" is turned into a single "_" - that'S completely wrong too :( There seems to be a probelm with the name sanitizer. I'll look into it -- Duesentrieb 13:08, 2 November 2006 (UTC)
Any news on this, I just saw it as well? Of course, replacing the plus sign with %2B works... jhawkinson 04:54, 27 March 2007 (UTC).

not recognising local duplicates at enwp?[edit]

e.g. [9] found: 10 KB, 300x300 pixels - not used - used on 1 pages [–] Number of the Beast - used on 1 pages [–] Arctic Monkeys different image (67 KB, 200x150 pixels) - used on 2 pages [–] דמיידה User:מיכאלי different image (7 KB, 130x191 pixels) - used on 1 pages [–] เล็บครุฑ (ภาพยนตร์)

but the enwp image is different, like the hewp and thwp ones. pfctdayelise 06:35, 4 November 2006 (UTC)

The toolserver still does not have a copy of the enwp database. I'm using query.php to check usage for enwp, but that doesn't give me enough details to detect duplicates. -- Duesentrieb 00:42, 5 November 2006 (UTC)

Und was ist mit dem &?[edit]

Hallo Duesentrieb, die oben gschilderten Probleme mit dem + im Dateinamen (macht man doch nicht) scheint es auch mit dem & zu geben - aber erst seit ein paar Wochen. Ja, inzwischen wird z.B. die Datei "Image:Germany coal mine Teutoburgia Schacht 2&1.jpg" nicht mal mehr in der Categorie als Bild angezeigt. Da das früher ohne Probleme war, vermute ich eine Verschlimmbesserung in der Software. & wat nu? Müsste wer die Datei umbenennen? -- 23:21, 6 November 2006 (UTC) PS: de:User:Arnoldius

commons:Image:Germany coal mine Teutoburgia Schacht 2&1.jpg wird nirgend korrekt angezeigt, nicht mal auf der Bildbeschreibungsseite. Vermutlich stirb die software bei dem versuch., das zu skalieren. Checkusage funktioniert jedenfalss korrekt für die Datei [10], allerdings nicht mit dem "tab" auf der Bildbeschreibungsseite. -- Duesentrieb 21:00, 8 November 2006 (UTC)

Fehlermeldungen im Bulk-Mode[edit]

Hi Duesentrieb, erst mal großes Lob und vielen Dank für dieses nützliche Tool.

Ich hatte heute leider Pech, im Bulk-Mode mit einer Liste von etwas mehr als 100 Bildern erschien erst eine Warnung, beginnend mit:

  • Warning: file_get_contents(h ttp:// ...

danach eine Meldung in rot:

  • Failed to fetch ...

und dann eine offenkundig unvollständige Liste. Mach ich was falsch? Gruß, --Mosmas 16:19, 22 November 2006 (UTC)

Du machst nichts falsch... ich verwende query.php für die englische Wikipedia, weil wir immer noch keine Datenbank-Kopie für diese Wiki auf dem toolserver haben (die ist einfach gigantisch groß, der Server hat sich bisher daran immer verschluckt). query.php beschränkt die mögliche Anzahl der Ergebnisse, und diese Beschränkung hat sich vor einiger Zeit geändert. Ich habe das Tool jetzt angepasst, der Fehler sollte in der nächsten Version nicht mehr auftreten. -- Duesentrieb 11:59, 27 November 2006 (UTC)

bug in recognizing image existance[edit]

Very handy tool. However there is a funny bug. If you search "Čipeva_sa_konjima.jpg" it will correctly identify the image used in 1 instance at hr.wikipedia, but then gives this message: "699 wikis searched. Čipeva_sa_konjima.jpg is not used anywhere!" This message directly contradicts the listing of that 1 instance! Please fix the error message. Thanks. CJLippert 04:01, 27 November 2006 (UTC)

CheckUsage detects the usage of files hosted on commons - commons:Image:Čipeva sa konjima.jpg does not exist, only a local copy at hr:Slika:Čipeva sa konjima.jpg exists. I agree however that the message wording is a bit strange in this context (its a bit tricky to provide internationalized messages for all possible outcomes) - there also should be a warning that the image doesn't exist on commons I guess. -- Duesentrieb 11:56, 27 November 2006 (UTC)

CheckUsage Error[edit]

CheckUsage returned error message since from 04.12.2006. Help please! --Kaganer 11:34, 7 December 2006 (UTC)

Still not working as of now. I receive the following error: "failed to get replication lag!" -- 20:23, 23 December 2006 (UTC)


Hi, when you check an image using the raw option, it fails on the english wikipedia, because illimit is too high for query.php Bryan 18:47, 15 January 2007 (UTC)

fixed a couple of days ago. -- Duesentrieb 11:30, 21 January 2007 (UTC)

pt.wikisource or wikiquote?[edit]

Hi, the CheckUsage is doubling the data from the pt.wikisource to the pt.wikiquote. See for example this:

  • - used on at least 100 pages [–]

Template:Categoria-anos (all ok here)

  • - used on at least 100 pages [–]

Template:Categoria-anos (the pt.wikiquote don't have a page with this exactly name)

555 01:31, 19 February 2007 (UTC)

Huh - the database view for pt.wikiquote actually points to the pt.wikisource database - so this bug effects all tools on the toolserver. I have asked DaB to fix it. Thanks for the info! -- Duesentrieb 10:44, 19 February 2007 (UTC)
should be fixed now. -- Duesentrieb 21:22, 19 February 2007 (UTC)

Checking usage of all photos per user[edit]

Could it be implemented - to find the most popular pics uploaded by particular commons user? --Monk 12:36, 19 February 2007 (UTC)

it could, but it's kind of hard on the server if a lot of people do it (and vanity searches are popular i guess). To do it manually, use the Gallery tool [11] to list all uploads (TSV mode would work best for that i guess), and chec them using the bulk mode of CheckUsage. -- Duesentrieb 21:24, 19 February 2007 (UTC)


This is in the footer at en.Wikinews. But it doesn't show up in CheckUsage. That is very worrying! pfctdayelise 00:18, 1 April 2007 (UTC)

Checkusage can only check things used on Wiki Pages, not things used by the skin, via CSS or JS, or on static portal pages. There'S no way to do that, because the database does not know about it. A workaround would be to collect all images used "internally" on a wiki page, so they are also used "on-wiki". -- Duesentrieb 11:30, 1 April 2007 (UTC)
I recommended it to Bawolff, and he did that.... but it still doesn't show up?? DB is up to date... what can it be? pfctdayelise 04:45, 2 April 2007 (UTC)
They just changed it to this one - very odd??? --pfctdayelise 07:05, 2 April 2007 (UTC)

There is currently a replication lag of 1 days, 4 hours, 45 minutes, 24 seconds for en.wikinews - "the database is up to date" is misleading, it only checks against de.wikipedia. Until recently, that was OK, because all databases except en.wikipedia where replicated from the same place, so they had the same replication lag. A few weeks ago, this was changed, so we now have three master databases, and thus three different replication lags, for en.wikipedia, 20 or so large projects (including de.wikipedia), and "the rest" (including en.wikinews). I should change the replag-display of the tools soon - wiki-specific tools already show the "correct" replication lag once you have selected a wiki, but CheckUsage should probably show all threee replag values, always. -- Duesentrieb 11:46, 2 April 2007 (UTC)

Ah right. Only good old replag. :) I remember the change. It will be hard to remember which wikis are on which clusters, so definitely they should all be listed I think. thanks. --pfctdayelise 01:11, 3 April 2007 (UTC)
I can't list them all, there are more than 800... -- Duesentrieb 13:54, 3 April 2007 (UTC)
But you can list the two smaller ones. E.g. 'En.wp lag: X minutes. Commons, ... lag: X minutes. All others: X minutes.' pfctdayelise 11:21, 4 April 2007 (UTC)
That would mean listing 20 wikis (the ones in "s2", including dewiki and commons), and maintaining that list somehow... :( -- Duesentrieb 11:42, 4 April 2007 (UTC)
Is that list going to change any time soon?? --pfctdayelise 00:19, 5 April 2007 (UTC)
BTW on the edit counters they say "dewiki (and 18 others)" and link to the email which lists all those wikis. CheckUsage should do something like that. it's important to know the lag, to properly judge if an image is being used or not... --pfctdayelise 08:00, 8 April 2007 (UTC)

English Wikipedia[edit]

The copy of the English Wikipedia used by CheckUsage is severely out of date. Is there any chance that you will be able to fix it? Andrewmackinnon 08:45, 16 June 2007 (UTC)

Ein "+" zwischen Portal und Diskussion[edit]

Hallo Duesentrieb,

dein Tool hat einen kleinen Bug. Aufmerksam darauf bin ich geworden, weil ein User auf Commons das Problem im Forum schilderte. Er zeigte das Problem an folgender Seite:

Sein Originaltext lautet:

Wer auf dieser Seite 
auf das angezeigte Portal klickt ..... kommt leider nicht an. Mutter Erde 15:55, 23 October 2007 (UTC)

Gruß, Körnerbrötchen 10:54, 24 October 2007 (UTC)

Parlez Français, ou d'autres langues[edit]

L'usage de Commons est difficile pour un francophone. Si Wikipédia est usuel dans plusieurs langues, pourquoi Commons est -il uniquement anglophone??? Il y en a marre de l'ominipotence de l'anglais...Jeveux donner de mes images aux utilisateurs Wiki mais je suis handicapé par la barrière de la langue! Damned !

Commons propose une traduction française de l'essentiel de son interface et de ses pages d'aide.
pourvu seulement que tu aies configuré ton navigateur en indiquant le français comme langue préférée, ce sera la langue affichée par défaut dans la majeure partie de l'interface.
De plus, si tu as un compte ouvert sur Commons, tu peux personnaliser un certain nombre d'options de localisation. 14:24, 4 December 2007 (UTC)

status snafu[edit]

Why does the CheckUsage tool say "Status: SNAFU" at the top? What is wrong with it? Mike R 16:15, 2 January 2008 (UTC)

Problem with CheckUsage[edit]

Hi, I've found a problem with your tool. For some of images it shows the error message:

"Fatal error: Allowed memory size of 16777216 bytes exhausted (tried to allocate 24 bytes)
in /home/daniel/public_html/WikiSense-live/common/WSInit.php on line 399"

Please see this example. I think that the error shows up for popular images, where there should be more results.

Regards, 19:37, 2 January 2008 (UTC)

CheckUsage error: undef func escapeHtml[edit]

I've got this error today:

Fatal error: Call to undefined function escapeHtml()
in /blah/blah/WSInit.php on line 418

when checking for Sparganium_erectum_section.jpg usage.
Regards, pl:CiaPan (talk) alias en:CiaPan (talk), 07:54, 25 January 2008 (UTC)

Replication lags[edit]

An image that I included in ro: four days ago still isn't listed, although the replication lag is reported less than two days? 00:27, 14 March 2008 (UTC)

I need help please if posible[edit]

I whould like if you explaned how I can insert a image (using gnu) when editing wiki. please reply by leaving a edit at the bottom of my user page, see here if you cant help please tell me who can thank you

Feature request[edit]

Is it possible to have all instances of an image being used on it's own page be either silently ignored or flagged with a little icon beside it or something? – Mike.lifeguard | @en.wb 16:03, 4 April 2008 (UTC)

"not found!not found!not found, but used!not found!not found!not found, but used!"[edit]

OMG ITZ NOT FOUND BUT USED!!!!!1!11!!eleven!!!

I know the image in question no longer exists, but still, isn't six run-on exclamations (in bold red text, too) just a bit too dramatic? :-) --Ilmari Karonen 13:34, 10 May 2008 (UTC)

Guck ma[edit]

Hi Düsentrieb, die Suchmaschine ignoriert Image:Vlad Mirita+Nico, May 2008.jpg: "766 wikis searched. Vlad_Mirita_Nico,_May_2008.jpg is not used anywhere!". Grüssle Mutti, Edle von AOL 08:21, 13 July 2008 (UTC)

Der Grund ist das "+" im Namen. Das sorgt leider für Verwirrung. -- Duesentrieb 20:41, 26 July 2008 (UTC)

Übel übel, sprach der Kübel. Die Suchmaschine erfindet übrigens auch Pluszeichen da, wo es eigentlich gar keine gibt. z.B. bei Image:Grabstelle (462) und 463 - Mutter Erde fec.jpg (Die Disk zum Berlin-Portal gibts natürlich und das Bild ist auch noch drin). Grüsse Mutti, Edle von AOL , heute via Hansenet 22:58, 11 August 2008 (UTC)

Bulk mode search question[edit]

When I search for (commons) Image:SouthHavenLightSouthPier.jpg‎ using the normal single image CheckUsage, it finds two matches on If I use the bulk mode search for


it finds neither of the pages, and returns one use for the first image. Am I doing something wrong here? Thanks, Parkerdr 01:52, 18 August 2008 (UTC)

Huh, that's odd. To be honest, I have no idea what'S going wrong. I'll try to look into it soon, but i'm a bit busy atm. -- Duesentrieb 07:05, 18 August 2008 (UTC)
no problem; it is kind of strange. Parkerdr 13:45, 18 August 2008 (UTC)



Great thanks for this tool. =) I just discovered the bulk mode, the number of wikis to search, and the "only articles" mode, and... it's wonderful! All that I searched since several months... =D

Delhovlyn[feel free to correct my bad English! This would be very useful to me.. =) ] 19:32, 28 September 2008 (UTC)

Problem with[edit]

Hi, I'm afraid there's a problem with CheckUsage feature:
A database error has occurred Query: SELECT * FROM itwiki_p.image WHERE img_name = 'Arnost_z_Pardubic_(Narodni_muzeum).jpg' Function: getImageInfo Error: 1146 Table 'itwiki_p.image' doesn't exist (sql-s2)
-- 16:08, 22 October 2008 (UTC)

I wonder why CheckUsage does need to perform "SELECT *" when it can just select only the data columns that it needs to query from the server. This would save server resources (and would speed up the scan), if the server was just sending only the data needed.
Well have not checked the sources of CheckUsage, but "SELECT *" does not mean "select all rows" (this "all" condition is the default in SQL, it is the "WHERE" clause or a max-rowcount boundary that can limit the rows returned) but really means « retreive all data columns in all the tables that referenced in the FROM clause and that match the WHERE restrictions and the boundary ». And in fact, for the scan performed by CheckUsage, it just needs to select a bounded number of page names that use an image, and does not need all the details about these pages, so it should NOT perform "SELECT *", but just "SELECT namespace,pagename" with a max-rowcount boundary (100 rows is much enough to allow displaying a list of pages from each wiki).
As a rule, I never use "SELECT *" in my programs, just because this creates a dependency with the current data model (where some datacolumsn ma be moved into other joined tables). A program normally needs only specific datacolumns, and the query can be modified when needed, but should not be dependant on the extra datacolumns that may be added in the future. The only place where "SELECT *" is useful is in existence subqueries (such as "SELECT columns FROM... WHERE EXISTS(SELECT * FROM...)" but even in this case, I prefer to write "SELECT columns FROM... WHERE EXISTS(SELECT 1 FROM...)" to help the SQL engine and really say it that I don't care about the data matched in the subquery. The only case where * is useful in SQL, is in "SELECT COUNT(*)" but even in this case it is completely equivalent to "SELECT COUNT(1)" because here also the content of datacolumns does not have any importance, what is needed is just the number of matching rows but none of the data in them.
As MediaWiki uses MySQL, and MySQL's internal query optimizer is not very advanced, it can perform too much work (within internal temporary storage tables or indexes), just to discard them at end. Don't allow it to do this extra work, notably when your SQL client does not need it. So remove "SELECT *" everywhere in your code and make sure that your selections are effectively selective (i.e. selective in the data rows to match, but also in the exact list of datacolumns you really need, and also by using bounded selections to force a limitation on the number of rows retreived). This is especially important for CheckUsage that is used by lot of people and that performs FAST queries on ALL wikis (except those for which it has its own local mirror made from database dumps). verdy_p 08:12, 10 December 2008 (UTC)

It'S open source [12], feel free to fork or provide patches. -- Duesentrieb 09:08, 10 December 2008 (UTC)

Wiktionary not scanned ?[edit]

Wiktionnaries also use some images in their pages (for illustrating their definitions). It's still not very frequent, but it starts being used. It seems that CheckUsage does not scan them. At least it should find the image usage made in French Wiktionary, because it takes almost all its images from Commons (very few images are stored locally except for the MediaWiki interface localisation and the local "Monobook" stylesheets and javascripts that generate images used implicitly in almost all pages). verdy_p 07:58, 10 December 2008 (UTC)

It does [13]. Unless you tell it to scan only wikipedias. -- Duesentrieb 09:07, 10 December 2008 (UTC)