Wiktionary:Grease pit

Wiktionary > Discussion rooms > Grease pit

Welcome to the Grease pit!

This is an area to complement the Beer parlour and Tea room. Its purpose is specifically for discussing the future development of the English Wiktionary, both as a dictionary and as a website.

The Grease pit is a place to discuss technical issues such as templates, CSS, JavaScript, the MediaWiki software, extensions to it, the toolserver, etc. It is also a place to think in non-technical ways about how to make the best free and open online dictionary of "all words in all languages".

It is said that while the classic beer parlour is a place for people from all walks of life to talk about politics, news, sports, and picking up chicks, the grease pit is a place for mechanics, engineers, and technicians to talk about nuts and bolts, engine overhauls, fancy paint jobs, lumpy cams, and fat exhausts. That may or may not make things clearer... Others have understood this page to explain the "how" of things, while the Beer parlour addresses the "why".

Permanent notice

  • Tips and tricks about customization or personalization of CSS and JS files are listed at WT:CUSTOM.
  • Other tips and tricks are at WT:TAT.
  • Everyone is encouraged to expand both pages, or to come up with more such stuff. Other known pages with "tips-n-tricks" are to be listed here as well.

Grease pit archives edit

August 2015

Category problemsEdit

This category and this one need to be updated. Thanks Codeofdusk (talk) 00:42, 1 August 2015 (UTC)

Categories update automatically, but sometimes it takes a while for the updates to show up. Purging the cache of a page or making a null edit will often update the category, too. —Aɴɢʀ (talk) 05:46, 1 August 2015 (UTC)

Eastern Mari code problemEdit

Here at Witkionary, the Eastern Mari language has the code "chm". At Wikipedia, however, its code is "mhr" (see, e.g., http://mhr.wikipedia.org). This creates an unsolvable problem: when I try to use templates like {{wikipedia}} or {{slim-wikipedia}} to link entries to the corresponding Eastern Mari Wikipedia articles, I get a Lua error message if I use the parameter lang=mhr, whereas, if I use lang=chm, I don't get an error message but the link is not to the Eastern Mari wikipedia (rather, it is to a non-existant namespace "Chm:" in the English Wikipedia). Given this situation, what's a well-intentioned average Wiktionarian to do?--Pereru (talk) 04:54, 3 August 2015 (UTC)

Module:wikimedia languages/data (I think I did that right- someone should check). DTLHS (talk) 04:57, 3 August 2015 (UTC)
I've just checked-- it still has the same problems. Maybe someone could check if that was the right change? --Pereru (talk) 01:09, 4 August 2015 (UTC)
It's only half of the change. By adding it to Module:wikimedia languages/data, you've specified that there's a Wikimedia code that should be turned into chm on Wiktionary. But there's also the reverse process, which is done in Module:languages/data3/c, to specify that the corresponding Wikimedia code for chm is mhr. —CodeCat 01:11, 4 August 2015 (UTC)
The template {{slim-wikipedia|lang=mhr}} gives an error while {{wikipedia|lang=mhr}} does not- why is this? DTLHS (talk) 01:13, 4 August 2015 (UTC)
Fixed. —CodeCat 01:21, 4 August 2015 (UTC)
OK, now {{slim-wikipedia|lang=mhr}} works, as e.g. in лум ‎(lum). But this creates a slightly irritating situation: one has to use lang=mhr with transwiki templates to get to the right Wikipedia, but lang=chm (or simply chm) with templates internal to Wiktionary. Why isn't there a single, cross-wiki code for Eastern Mari? Why not mhr also for Wiktionary? Was this just a mistake made at some point, or is there a deeper reason for the two codes? --Pereru (talk) 19:12, 4 August 2015 (UTC)
You need to edit Module:languages/data3/c to add mhr to chm (I can't do it). DTLHS (talk) 19:15, 4 August 2015 (UTC)
[1]. - -sche (discuss) 20:03, 4 August 2015 (UTC)
Was this supposed to make {{slim-wikipedia|lang=chm}} link to the mhr Wikipedia? Because this is still not working -- in лум, I used {{slim-wikipedia|lang=chm}}, and it still doesn't link to the mhr Wikipedia -- it only does if I use {{slim-wikipedia|lang=mhr}}. --Pereru (talk) 07:01, 5 August 2015 (UTC)
Yes, that's what it was intended to do. If it didn't work, I suspect the problem may be with {{slim-wikipedia}}. (Perhaps its guts haven't been kept as up-to-date as the usual templates'.) - -sche (discuss) 16:50, 5 August 2015 (UTC)
The problem exists with {{wikipedia}} too, and has to do with how Module:wikimedia languages's getByCodeWithFallback function works. What currently happens in these cases is:
  1. Look up "chm" as a Wikimedia language code.
  2. If one is found, return that.
  3. If none is found, look for a regular language with that code.
  4. If one is found, return its language code.
Currently, step 4 is the problem. The language's Wiktionary code gets returned, which may not be a valid Wikimedia code. getWikimediaLanguages would return a valid Wikimedia language for the language, if one is available. The difficulty is that getWikimediaLanguages returns not one language, but a list of them. This is to account for languages like Bosnian, Croatian, Serbian which we all merge into Serbo-Croatian. So which item on the list should be selected here? The question "what is the Wikimedia code belonging to Wiktionary language x?" doesn't have a single defined answer. A possible solution could be to select the first item if that's the only one in the list, and show an error that the code is ambiguous otherwise. —CodeCat 18:23, 5 August 2015 (UTC)
Or we could do the opposite. Have {{wikipedia}} take a Wikimedia code and deduce the language code from that. --WikiTiki89 18:30, 5 August 2015 (UTC)
That's what it does currently, but Pereru doesn't like that. The fallback is just there for when someone uses "cmn" as the code, since Wikimedia calls that "zh". —CodeCat 18:33, 5 August 2015 (UTC)
Ok, looking at it now I'm starting to get quite confused by why this doesn't work. {{#invoke:wikimedia languages/templates|getByCodeWithFallback|cmn|getCode}} > zh, but {{#invoke:wikimedia languages/templates|getByCodeWithFallback|chm|getCode}} > chm. Why is the first returning the right code, but the second isn't? —CodeCat 18:43, 5 August 2015 (UTC)
I found the source of the problem. On line 40 of Module:wikimedia languages, the code if not mw.language.isKnownLanguageTag(code) then checks if a given code is recognised by the software as a valid interwiki code. For "cmn" this returns false, which it should as there are no cmn: sites. But for "chm" it returns true, indicating that this code is valid for interwikis. So it's a problem with the wiki's own configuration. —CodeCat 18:51, 5 August 2015 (UTC)
But you still haven't answered Pereru's earlier question: Why do we use chm, which is the general code for Eastern and Western Mari, to represent Eastern Mari instead of mhr, which is meant to represent only Eastern Mari? --WikiTiki89 18:47, 5 August 2015 (UTC)
Ask User:-sche, who made the change in diff back in 2013. —CodeCat 18:53, 5 August 2015 (UTC)
I was following Anatoli's suggestion here and here that "With Mari, I would rather delete mhr and leave the name "Mari". Standard Mari is "Eastern Mari" or "Meadow Mari" and chm is more common." I wasn't aware that it had a Wikipedia, or I would have pointed out that we should just use the assigned Eastern Mari code. We could (should) just switch now... - -sche (discuss) 21:38, 5 August 2015 (UTC)
It doesn't really matter, which we choose. Mari (chm) is a political term but linguistically, it makes more sense to use both "Eastern Mari" and "Western Mari". When they use just "Mari", "Eastern (Meadow) Mari" is implied. There is a Mari Wikipedia but the language used there is "Eastern Mari". --Anatoli T. (обсудить/вклад) 11:03, 6 August 2015 (UTC)
The point is not the name but the code. Given that Eastern Mari has its own code, and that its Wikipedia uses that code, we could avoid all the contortion described in this thread if we just used Eastern Mari's code for Eastern Mari, rather than using the Mari macrolanguage's code for Eastern Mari like we do now. - -sche (discuss) 21:48, 6 August 2015 (UTC)

Thai transliterations with tonesEdit

Discussion moved to Wiktionary:Beer parlour/2015/August#Thai transliterations with tones.

Strange difference between entry and previewEdit

When I look at Eastern Mari пӧрт, the third-person form in the possessed forms table is shown as "пӧрты", which is incorrect: it should be "пӧртшӧ". Strangely, when I edit and preview this page, the correct form "пӧртшӧ" shows up in the table (see [2]). Why the difference? Is it because I just edited the possessed forms template, and some time is necessary for the change to be processed? --Pereru (talk) 07:05, 5 August 2015 (UTC)

When you edit a template, the change doesn't immediately propagate to all of the entries that use it. The system has to run a separate process to update them all, and if other changes have been made to other templates, your changes may have to wait for the others to be done first. In a few cases involving hundreds of thousands of transclusions, I've seen it take weeks. When you click on Edit, the system updates everything while preparing the edit window- so you can force all the changes to be applied by simply clicking on Edit and then clicking on Save without making any changes. This is called a null edit, and it won't show on the revision history. Chuck Entz (talk) 07:59, 5 August 2015 (UTC)
BTW sometimes it never seems to properly update a category unless you force things with null edits. Benwing (talk) 08:22, 5 August 2015 (UTC)
Categories do get updated, just much more slowly. --WikiTiki89 11:15, 5 August 2015 (UTC)
Even edited entries can take days to appear in a new category to which they have been assigned by a template, say, by a switch. DCDuring TALK 13:18, 5 August 2015 (UTC)

Maybe Cyrilic vs. Latin script, or something else?Edit

In Eastern Mari пӧртeм, the first form in the declension table is also "пӧртeм", i.e., the nominative singular form. However, when I click on it, it creates a new "пӧртeм" page, as if there wasn't one already. At first I thought that the problem was with the "ӧ" character: there is one in both the Latin and the Cyrilic script, and apparently they have different codes. But when I tried changing this character, the situation was still the same: clicking on the nominative singular form "пӧртeм" in the table still creates a new "пӧртeм" page. What the heck is going on here? (NB: in the main lemma page пӧрт, this problem doesn't exist: there are two inflection tables, in the second of them there is a form "пӧрт"; this form links back to the same page, i.e. if I click on it no new page is created.) --Pereru (talk) 11:35, 5 August 2015 (UTC)

If you press Ctrl+F in your browser and search for a Latin-script "e", you will see that you used the Latin-script "e" in the Cyrillic word that should be пӧртем ‎(pörtem). Ctrl+F is a good way to find these problems. --WikiTiki89 11:47, 5 August 2015 (UTC)
Wow, that works. Thanks for the tip! --Pereru (talk) 12:02, 5 August 2015 (UTC)

Category:Chinese terms with uncreated formsEdit

When using Template:zh-new, sometimes an entry will appear in Category:Chinese terms with uncreated forms even when it has no reason to. An entry can be removed with a null edit, but the point is that some entries have no reason to appear here:

  • 番泻叶's traditional form 番瀉葉 does not exist yet so it belongs in this category
  • 傲嬌 and 傲娇 both exist so the inclusion in this category is odd
  • 梗阻 doesn't even have any alternate forms at all so its presence in the category is exceedingly odd

suzukaze (tc) 08:53, 6 August 2015 (UTC)

JavaScript broken?Edit

Is it just me or did all the JavaScript stop working suddenly? --WikiTiki89 23:29, 6 August 2015 (UTC)

And now it's back... False alarm, or did someone just fix it? --WikiTiki89 23:30, 6 August 2015 (UTC)
It’s been unstable for a couple of days. — Ungoliant (falai) 23:39, 6 August 2015 (UTC)

This Wednesday, August 5, for hours I couldn't load Wiktionary because of JavaScript, the actual content blinked for a second then the screen went completely blank and permanently loading. When I disabled JavaScript on my browser (Firefox 39.0 on Windows 7) Wiktionary loaded normally (i.e., naturally, as normally as possible, without the benefit of any of our JS gadgets). --Daniel Carrero (talk) 06:32, 7 August 2015 (UTC)

Incomplete transwikiEdit

The transwiki from Wikipedia to Wiktionary of Appendix:List of Proto-Indo-European roots was incomplete, partly due to a cut and paste move and partly because the page import simply didn't capture all the relevant edits, as I described in my message on the appendix's talk page. Is there anything that realistically could or should be done about this now on Wiktionary's end, besides my message there? I'd understand if there weren't any other sensible courses of action, due to the number of revisions involved. Graham87 (talk) 12:47, 7 August 2015 (UTC)

Built in Lua editorEdit

At one point editing Module: pages brought up an editor that automatically did syntax highlighting and tabs. Now it's just like any other page. Is there some preference that needs to be checked to get this back? DTLHS (talk) 20:34, 7 August 2015 (UTC)

There's a button that looks like "<>" on the top left corner that switches between the plain text editor and the fancy code editor. --WikiTiki89 20:36, 7 August 2015 (UTC)
The top left corner of what? Should there be a button here? DTLHS (talk) 20:41, 7 August 2015 (UTC)
Oh, it looks like your JavaScript failed to load or crashed or something (which is why the whole toolbar where the button would have been is missing). Try a cache-bypassing refresh (usually Ctrl+Shift+R). --WikiTiki89 20:44, 7 August 2015 (UTC)
Yeah that didn't help- I can get it to work in Chrome but not Firefox. It only works if I log out. So is there a preference I checked to somehow turn it off? DTLHS (talk) 20:48, 7 August 2015 (UTC)
No this is a JavaScript error. I had a similar problem yesterday (see above). You can open the JavaScript console (Ctrl+Shift+J in Chrome) and see if it shows any errors (you might need to refresh with the console open to get the errors to actually show up in the console). Otherwise you're having the same problem I had where the JavaScript just failed to load, and all you can do is just keep refreshing till it works. --WikiTiki89 21:06, 7 August 2015 (UTC)
I haven't seen the editor for at least 6 months, so I guess I'm just screwed. DTLHS (talk) 21:16, 7 August 2015 (UTC)
Does other JavaScript work for you? Do you ever see the toolbar above the edit window? --WikiTiki89 21:18, 7 August 2015 (UTC)
I had unchecked Show the edit toolbar (requires JavaScript) and Enable enhanced editing toolbar in preferences- thanks for the help. DTLHS (talk) 21:21, 7 August 2015 (UTC)


Russian terms with |notrcat=1 shouldn't get into Category:Russian terms needing accents. E.g. козел ‎(kozjól) shouldn't have a stress. The accented form is козёл ‎(kozjól) --Anatoli T. (обсудить/вклад) 13:43, 9 August 2015 (UTC)

Sajax library will be removed from MediaWiki on 12 AugustEdit

As announced on en.WP on 6 August, the Sajax library will be removed from MediaWiki on 12 August and any scripts which use this library (on en.WP, there were 11,000 users with such scripts) will break. Among other things, the discussion on 'pedia suggests that it will no longer be possible to use document.write and document.writeln. A cursory search of this site for those phrases suggests that ~100 users (of whom about a dozen have been active lately) may be affected, mostly users of the Monobook skin but also apparently e.g. @ZxxZxxZ's User:ZxxZxxZ/common.js. - -sche (discuss) 19:57, 9 August 2015 (UTC)

Will affected users be alerted or will things just break silently? Equinox 20:03, 9 August 2015 (UTC)
I'm not aware of a plan by WMF to alert users. Mr. Stradivarius mentioned that he'd try to mass-message users, though whether he meant globally or just on en.WP I'm not sure. The list he compiled of affected scripts found only eight (fewer than I found) on this wiki, to wit:
enwiktionary        User:Cirt/Gadget-HotCat.js
enwiktionary        User:Conrad.Irwin/edittools.js
enwiktionary        User:Daniel Carrero/Gadget-HotCat.js
enwiktionary        User:Eirikr/edittools.js
enwiktionary        User:Mike Dillon/Scripts/topicCats.js
enwiktionary        User:Prillen/monobook.js
enwiktionary        User:Sack36/edittools.js
enwiktionary        User:Stephen G. Brown/edittools.js
- -sche (discuss) 20:10, 9 August 2015 (UTC)
edittools.js has now stopped working. Can anything be done? @ZxxZxxZ, is it working for you? --Vahag (talk) 13:06, 16 August 2015 (UTC)
It's working for me. --Z 15:00, 16 August 2015 (UTC)
  • Yes, it is working now, thanks. --Vahag (talk) 07:04, 27 August 2015 (UTC)
A phabricator tiket which Eirikr helpfully found and linked to here says that most scripts which were broken by the removal of Sajax can be fixed with a few simple copy-and-paste replacements. The only en.Wikt regulars who have scripts that use the Sajax phrases mentioned in that phabricator ticket are User:Stephen G. Brown in User:Stephen G. Brown/edittools.js and User:Daniel Carrero in User:Daniel Carrero/Gadget-HotCat.js: if either of you finds that your tool no longer works, replace "sajax_init_object(" with "new XMLHttpRequest(" and see if that fixes it. - -sche (discuss) 02:40, 27 August 2015 (UTC)

Looking at the best way to link quotes to works at enWSEdit

Hi. Just was looking at the best way (preferred means?) to link to enWS in the use of {{quote-book}}. In this example I used interwiki rather than a url parameter, so if that is not preferred then please let me know. I regularly trip across archaic and less used words in these older books, and if there was a particular cite/quote template that was configured for easy use of the Wikisources, then I try to get around to use it. — billinghurst sDrewth 06:09, 10 August 2015 (UTC) (please ping me if there is a reply)

Automatic pronunciation (IPA and hyphenation) for EsperantoEdit

Discussion moved from Wiktionary talk:Main Page.

The pronunciation and stress rules of Esperanto are completely regular. Maybe we could create a template/script for the pronunciation, which automatically inserts IPA and hyphenation which can be used in all Esperanto lemmas. Just like the header and conjugation are now used everywhere. Robin van der Vliet (talk) (contributions) 15:46, 10 August 2015 (UTC)

@Robin0van0der0vliet Module:eo-IPA. Needs more test cases. DTLHS (talk) 18:23, 10 August 2015 (UTC)
@DTLHS I added a few test cases, but my test cases also contain stress marks, it would be great if your script could also add them. The stress rules are luckily very easy. Stress is always on the penultimate syllable. There is only one exception to this rule, and this is when the final -o sound is elided and replaced with an -'.

Some examples:
[aˈt͡ʃeti] aĉeti
[alkoˈhola] alkohola
[eu̯ˈropo] Eŭropo
[faˈt͡sila] facila
[hoˈnoro] honoro
[famiˈlio] familio
[famiˈli] famili'

I noticed that some articles used dots to separate the syllables. Do you think this should also happen? I did this in my example to clarify the stress boundaries. Most articles don't do this, probably better to include hyphenation for this.
Robin van der Vliet (talk) (contributions) 20:51, 10 August 2015 (UTC)
How are syllables divided? DTLHS (talk) 21:27, 10 August 2015 (UTC)
Syllable structure in Esperanto is of the form (s/ŝ)(C)(C)V(C)(C). More information about Esperanto phonotactics is here.
In alkohola for example, how would you decide between al-ko-hol-a and alk-o-hol-a, since both seem to follow the rules? DTLHS (talk) 00:55, 11 August 2015 (UTC)
I'd actually guess al-ko-ho-la. Most languages put a single consonant between vowels with the following syllable, and split two-consonant clusters (English is the weird language out here). Benwing (talk) 01:51, 11 August 2015 (UTC)
I would actually guess that Esperanto is not so rigidly prescribed that syllabification is part of it. Presumably people with different backgrounds might syllabify differently, and this wouldn't cause difficulties. —CodeCat 01:54, 11 August 2015 (UTC)
If that's the case then I guess a pronunciation module isn't possible. DTLHS (talk) 02:19, 11 August 2015 (UTC)
You could just ignore the syllabification and use the ugly but easy strategy of putting the stress mark immediately before the vowel. --WikiTiki89 02:23, 11 August 2015 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── There are syllabification rules for Esperanto given here: [3] You may as well follow them. Benwing (talk) 02:27, 11 August 2015 (UTC)

Thank you, I'll try to implement that. DTLHS (talk) 02:31, 11 August 2015 (UTC)
@Robin0van0der0vliet Added a syllabification function, please check new test cases. DTLHS (talk) 05:10, 11 August 2015 (UTC)
@DTLHS It looks pretty good now. But it goes wrong when there are 2 vowels following each other. Syllables always contain one vowel or diphthong, never more or less.
The module also gives an error if you try to get the IPA of a monosyllabic word (such as "mi" which should return "[mi]"). Those words don't have stress.
Robin van der Vliet (talk) (contributions) 11:29, 11 August 2015 (UTC)
@Robin0van0der0vliet Fixed the remaining errors. DTLHS (talk) 15:47, 11 August 2015 (UTC)
Two more questions: should the template note the syllable breaks, as on biokomputiko? And should it use [] or //? DTLHS (talk) 16:01, 11 August 2015 (UTC)
I don't think we should have syllable breaks in the IPA, as most articles list this separately in the hyphenation. And I also think // would be better now, as I see most lemmas use this method. Robin van der Vliet (talk) (contributions) 17:31, 11 August 2015 (UTC)
Hyphenation is not the same thing as syllable breaks: hyphenation applies to the written language and and syllable breaks to the spoken language. --WikiTiki89 17:51, 11 August 2015 (UTC)
@Wikitiki89 Could you give me an example of a difference? I couldn't think of a difference in Esperanto. Robin van der Vliet (talk) (contributions) 21:00, 11 August 2015 (UTC)
As a construction language, Esperanto might not have any differences. Although if you study the actual speech of Esperanto speakers to determine their syllabification, you might find it does not match the standard. --WikiTiki89 03:10, 12 August 2015 (UTC)
@Robin0van0der0vliet I've analyzed all of the Esperanto IPA on Wiktionary and compared it with the module. Anything on that list must either have the wrong IPA, some rule the module isn't accounting for currently, or some irregularity that can't be programmed. DTLHS (talk) 17:43, 12 August 2015 (UTC)
@DTLHS There are a few things that are not covered by the module, I see now.

Consonant and the letter ŭ are pronounced differently on its own if you are talking about the letter. (example)
"b" - currently: /b/ - should be: /bo/
"ĥ" - currently: /x/ - should be: /xo/
"ĉ" - currently: /t͡ʃ/ - should be: /t͡ʃo/
"ŭ" - currently: /u̯/ - should be: /u̯o/
"a" - currently correct: /a/
"u" - currently correct: /u/

And there are also a few words on your list that are actually multiple words:
"Rio-de-Ĵanejro" - currently: /riodeʒaˈnei̯ro/ - should be: /ˈrio de ʒaˈnei̯ro/
"kiel vi fartas" - currently: /kielviˈfartas/ - should be: /ˈkiel vi ˈfartas/
"eventa horizonto" - currently: /eventahoriˈzonto/- should be: /eˈventa horiˈzonto/

I also noticed that a few words in that list use allophones. Sounds that are acceptable in Esperanto, but are not the standard form. (source)
Robin van der Vliet (talk) (contributions) 00:40, 13 August 2015 (UTC)
Thanks. Could you add test cases for any mistakes you see? DTLHS (talk) 00:47, 13 August 2015 (UTC)
@DTLHS I added a few more test cases. I fixed the test case for T-ĉemizo already and I fixed the test case for k.t.p. almost.
(Sorry for the late response by the way, I was on holiday in Luxembourg.)
Robin van der Vliet (talk) (contributions) 10:20, 27 August 2015 (UTC)
@DTLHS Maybe you could also create a function which returns the last part of a word, so we could easily categorize words that rhyme in Esperanto.
It should only return the part (in IPA) after the stress and begin with the first vowel.
Word IPA Rhymes
knabo (boy) /ˈknabo/ abo
frata (brotherly) /ˈfrata/ ata
Afganistano (Afghanistan) /afɡanisˈtano/ ano
aliancano (ally) /alianˈt͡sano/ ano
ekspediciano (expedition member) /ekspedit͡siˈano/ ano
Robin van der Vliet (talk) (contributions) 18:38, 31 August 2015 (UTC)
@Robin van der Vliet I've finished adding this functionality to the module, although I think you'll find that there are certain editors here who are vocally opposed to any kind of automatic rhyme categories. DTLHS (talk) 17:16, 2 September 2015 (UTC)
@DTLHS Thank you! Why would anybody oppose it, if it works perfectly? Currently IPA, hyphenation and rhyming templates are already rare for Esperanto.
I think that one of the last few steps is now creating a template which inserts IPA, hyphenation and rhyming templates and which has parameters for overrides (for exceptional circumstances) and an optional parameter for a pronunciation file.
Robin van der Vliet (talk) (contributions) 17:56, 2 September 2015 (UTC)
@Robin van der Vliet {{eo-pron}}, ĉevalido. DTLHS (talk) 22:17, 3 September 2015 (UTC)
@DTLHS Great, thank you! It looks pretty much finished now. What now? Should we add it to more pages manually (or even automatically)? Robin van der Vliet (talk) (contributions) 22:33, 3 September 2015 (UTC)
You should solicit opinions of other active Esperanto editors first. DTLHS (talk) 22:39, 3 September 2015 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── @Amire80, Embryomystic, Mr. Granger, Mutichou, Pereru, Robin Lionheart, Stephen MUFC, Tezamen, Tn4196 Pinging a few active Esperanto editors. Could some of you take a look at the newly created {{eo-pron}} and give feedback on it? This template automatically generates IPA, hyphenation and rhymes for use in Esperanto entries. Most Esperanto entries currently don't even have a pronunciation section at all. Robin van der Vliet (talk) (contributions) 23:35, 4 September 2015 (UTC)

Sorry for the delay in responding; the ping didn't work for some reason. I have a few thoughts.
  1. I imagine that for a small number of entries, we may still have to do pronunciation manually. It might be tricky to make the template work for both -j and j, for example.
  2. The template currently uses /i̯/ and /u̯/ where I would use /j/ and /w/ (i.e., for syllable-final j and ŭ). Is there any reason for using the symbol instead of j, aside from the fact that that's what Wikipedia does? It seems preferable to use the same symbol to transcribe the letter j wherever it appears—both L. L. Zamenhof and phonetician John C. Wells say that Esperanto orthography is phonemic ("one letter, one sound").
  3. Regardless of what we decide about i̯/u̯/j/w, the word ludejo should be transcribed with a j, not a i̯. Likewise with necesejo, kajo, vojo, etc.
  4. Regardless of what we decide about i̯/u̯/j/w, there are a small number of words in which ŭ must be transcribed as /w/ rather than /u̯/—words where ŭ occurs syllable-initially, such as ŭato, ŭaŭ, and ŭonbulismo. (This might be rare enough that it would be better to just do those entries manually.)
  5. The current version of the template produces the hyphenation "lu‧nlu‧mon" for lunlumon, which strikes me as no good. I would hyphenate it "lun‧lu‧mon".
Overall, the template looks great—I look forward to seeing it added to entries. —Mr. Granger (talkcontribs) 19:43, 7 September 2015 (UTC)
@Mr. Granger I added some test cases for these concerns. For lunlumon, the book that was linked earlier states that "a consonant followed by l or r goes with the l or r", but this seems to produce poor results in this case. Is there some additional rule that can be followed? DTLHS (talk) 19:51, 7 September 2015 (UTC)
I'm guessing that the consonants l, m, n, and r are exceptions to that rule. There might be other exceptions to the rule too—in the hypothetical compound word Ĉeĥlando ("Czechland"?), I have to think there would be a break between the ĥ and the l.
That book also says that words are usually split between two roots and between a root and an affix, which seems accurate, and would generally solve this problem, because consonant clusters like "ll", "ml", "nl", and "rl" usually occur in compound words. (It would also deal with vesperstelo, whose hyphenation "ves‧pers‧te‧lo" also strikes me as somewhat odd.) I don't know if there's an easy way to implement this rule in the template, though.
Here's one idea: Would it be possible to modify the template to allow users to input boundaries between roots and affixes when necessary? I'm imagining something like {{eo-pron|vesper|stelo}}. Then the template could take that into account when determining hyphenation. —Mr. Granger (talkcontribs) 20:17, 7 September 2015 (UTC)
@Mr. Granger, @Robin van der Vliet I've updated the module so you can do things like {{eo-pron|vesper|stelo}} (any artificial syllable break- the module will split the remaining syllables according to the usual rules). DTLHS (talk) 19:12, 9 September 2015 (UTC)
Could you explain in what circumstances words should be transcribed with j, not a i̯ (ludejo)? DTLHS (talk) 19:12, 9 September 2015 (UTC)
Thanks! Regarding the letter "j", in ludejo, vojo, etc., the "j" is a consonant at the beginning of a syllable (per the syllabification rules described in the book linked above), so it should certainly be transcribed as /j/. To put it simply, whenever "j" is immediately followed by a vowel, it should be transcribed as /j/. (As I mentioned above, I would transcribe the letter "j" as /j/ no matter where it appears, but it should certainly be transcribed as /j/ in this context.) —Mr. Granger (talkcontribs) 22:49, 9 September 2015 (UTC)
I see. In that case the problem is that the module will split ludejo into letters as l-u-d-ej-o, instead of l-u-d-e-j-o, since it looks for digraphs first. I'll have to think about how to fix that. DTLHS (talk) 00:58, 10 September 2015 (UTC)
Another issue: the template currently transcribes poŭpo as /poˈwpo/, when it should be either /ˈpou̯po/ (in the system the template currently uses) or /ˈpowpo/ (as I would transcribe it). —Mr. Granger (talkcontribs) 01:26, 10 September 2015 (UTC)
How would you transcribe ŭaŭ? DTLHS (talk) 01:34, 10 September 2015 (UTC)
/waw/. I would transcribe "ŭ" as /w/ wherever it appears. —Mr. Granger (talkcontribs) 01:51, 10 September 2015 (UTC)
@Mr. Granger I've removed the wikipedia-based transcriptions, so everything should be transcribed as j now. Makes it a lot simpler. Could you go over the testcases again to make sure all the expected values are correct? DTLHS (talk) 20:35, 10 September 2015 (UTC)
ŭ should be /wo/ and Eŭropo should be /ewˈropo/, for consistency. Other than that, they look good.
I've also noticed one more mistake—the template currently transcribes ankaŭ as /anˈkaw/, but it should be /ˈankaw/. —Mr. Granger (talkcontribs) 21:23, 10 September 2015 (UTC)
Also, I think "j" should be added to the list of exceptions to the syllabification rule about l and r. So najlo should be hyphenated naj‧lo instead of na‧jlo. —Mr. Granger (talkcontribs) 21:36, 10 September 2015 (UTC)
Got it. Now passing all tests (except for compound words). DTLHS (talk) 21:44, 10 September 2015 (UTC)
Wonderful! —Mr. Granger (talkcontribs) 21:47, 10 September 2015 (UTC)
  • @DTLHS: I've found another issue. The template currently hyphenates antaŭen as an‧ta‧ŭen, when it should be an‧taŭ‧en. When "ŭ" occurs between two vowels, it should be in the same syllable as the preceding vowel, not the following vowel. (This is a difference between "ŭ" and "j", because when "j" occurs between two vowels, it should be in the same syllable as the following vowel.) —Mr. Granger (talkcontribs) 16:56, 12 September 2015 (UTC)
Fixed. DTLHS (talk) 17:36, 12 September 2015 (UTC)
Thanks! Now there's a problem with antaŭiri – it's transcribed /anˈtawiri/ when it should be /antawˈiri/. —Mr. Granger (talkcontribs) 17:46, 12 September 2015 (UTC)
Also fixed. DTLHS (talk) 17:58, 12 September 2015 (UTC)
antaŭa seems to have a similar problem—it's being transcribed /ˈantawa/ instead of /anˈtawa/. Likewise antaŭu and antaŭo. —Mr. Granger (talkcontribs) 18:32, 12 September 2015 (UTC)

TweenkBot discussionEdit

I want to request the bot flag for User:TweenkBot for these tasks:

  • Adding animacy specifier to Polish masculine nouns (converting 'm' gender to 'm-in', 'm-an' or 'm-pr). An advanced student may look at this and figure out the inflected forms without consulting the full declension table.
  • Removing the Turkish template spam, such as Template:second-person singular imperative of.

I would be using pywikibot's replace.py for these tasks. In the first case, I use an interactive replacement function that prompts me for each noun, while in the second one I just use the regex functionality. --Tweenk (talk) 00:04, 11 August 2015 (UTC)

Per WT:BOT, I think you should create a vote. I don't mind your edits too much on your main account, but it would be better if you requested a flood flag when you're about to run them until you get a bot approved. —Μετάknowledgediscuss/deeds 00:08, 11 August 2015 (UTC)
Just a note — please don't run them on the bot account yet, because that account isn't autopatrolled. Thanks —Μετάknowledgediscuss/deeds 00:10, 11 August 2015 (UTC)
Let's just whitelist the bot (for any manually-reviewed edits; for fully automated edits it still needs a WT:BOT vote). - -sche (discuss) 02:18, 11 August 2015 (UTC)
I created the vote some time ago: Wiktionary:Votes#User:TweenkBot_for_bot_status
Do I also need to announce it somewhere? --Tweenk (talk) 14:12, 17 August 2015 (UTC)

German nouns which have all three gendersEdit

We categorize entries by unusual features like use of Æ, and by inflectional features like uncountability. I think it might be useful to categorize German nouns that can have all three genders. (What about nouns that have two genders? Should 3- and 2-gender nouns be distinguished? Presumably the categories could be added automatically by {{de-noun}} based on whether or noun g3= was present.) According to templatetiger there are six nouns with 3 genders: Dschungel, Jogurt/Joghurt, Wimpel, Lahmacun, Nutella. An alternative to a category, feasible if the number of 3-gender nouns remains very small, is to manually update the usage notes currently in Joghurt, Wimpel, etc (perhaps templatizing the part that talks about which other nouns have three genders, so it can be updated centrally). - -sche (discuss) 23:03, 15 August 2015 (UTC)

There are 96 entries which use g2=, but several of these are erroneous, e.g. Rothaarige, which needs to be split to have two ===Noun=== headers. - -sche (discuss) 23:11, 15 August 2015 (UTC)
For maintenance purposes or checking our completeness against a list of such nouns, it is failrly easy to use CirrusSearch with a regular expression to find g2 and g3 parameters. It does take some skill to get the regex right, but even I have done some simple cases. DCDuring TALK 23:34, 15 August 2015 (UTC)
For determining which entries do or don't use a given parameter of a given template one can simply use templatetiger (as I did above). I just wondered if we wanted a category which would automatically keep up-to-date with new entries, if we thought that "which nouns have all three genders" was something users might be interested in (I think it is), like we apparently (rightly, IMO) think countability is something people are interested in since we categorize by that. - -sche (discuss) 23:41, 15 August 2015 (UTC)
Un/countability has direct consequences for grammatical speech and writing, which warrants the labels. At the time of debate about the category we didn't have Cirrus Search AFAICR. The un/countable categories are too big for most purposes other than counting them, which doesn't have much point either.
Categories that are merely interesting and not too big could be in an Appendix. Perhaps the 2- and 3-gender nouns are of that type. Is there much benefit to the real-time capability? DCDuring TALK 00:10, 16 August 2015 (UTC)

Adding a new POS for poscatboilerEdit

The documentation is very confusing, and I don't know where I'm supposed to add 'Relational' so that Category:Pipil relationals won't throw an error. (Also, it's never been used as an L3 header before AFAIK, so maybe it needs to be added to something else as well?) —Μετάknowledgediscuss/deeds 17:29, 16 August 2015 (UTC)

You would need to edit Module:category tree/poscatboiler/data/lemmas. Why "Relational" as opposed to "Relational noun"? DTLHS (talk) 19:15, 16 August 2015 (UTC)
What are "relationals" anyway? The ones I looked at in that category all look like postpositions to me. —Aɴɢʀ (talk) 06:20, 19 August 2015 (UTC)

Noun plural formsEdit

The entry fuckbags, among others, is categorised in Category:English noun plural forms. I'm assuming there's a bug somewhere in the WT:ACCEL tool --A230rjfowe (talk) 09:54, 17 August 2015 (UTC)

No, it's the intentional result of standardizing how plural forms are treated across languages; see Wiktionary:Beer parlour/2015/August#Category:.28langname.29_plurals_and_Category:.28langname.29_noun_plural_forms. Even in English, "plurals" is a bit ambiguous, because other parts of speech besides nouns have plural forms: the most widespread and native are verbs (he is, they are), but also some French-derived adjectives (manqué). - -sche (discuss) 17:34, 17 August 2015 (UTC)

Category:Plurals with a red link for singularEdit

...contains entries like Appendix:Proto-Algonquian/-hpanali, although its singular is a blue link. Why? - -sche (discuss) 03:50, 19 August 2015 (UTC)

I'd imagine, because *-hpani doesn't exist, and that's what the code's checking for. Having stuff like {{new ca plural}} in there is irritating also. Renard Migrant (talk) 16:48, 20 August 2015 (UTC)

New feature "Watch changes in category membership"Edit

recent changes page view with categorization

Hi, coming with this week’s software changes, it will be possible to watch when something is added to or removed from a category (T9148). The feature has been requested by the German Community and is part of the Top 20 technical wishlist. The feature was already deployed to Mediawiki.org on August 18 and it will be rolled out on Wiktionary between 6-8 pm UTC today. It will be available on all Wikipedias from Thursday 20 on, likewise between 6-8 pm UTC. In this RFC-Proposal, you can find the details of the technical implementation. The feature was implemented via a new "recent changes" type for categorization. Through this, categorization will be logged and shown on the recent changes page. The categorization logg in "recent changes" is the data base for the watchlist: When you watch a category, added or removed pages from that category will be shown on the watchlist. The categorization of pages can be turned off in the watchlist preferences as well as recent changes preferences. If you have any questions or remarks about the feature or if you find a bug, please get in touch! Bugs can also be reported directly in Phabricator, just add the project “TCB-Team” to the respective task. Cheers, Birgit Müller (WMDE) (talk) 14:34, 19 August 2015 (UTC)

This is an amazingly useful change and I'm glad it has been added! —CodeCat 14:40, 19 August 2015 (UTC)
I can't wait! --WikiTiki89 14:51, 19 August 2015 (UTC)
Sounds good. Can I turn off everything in recent changes except for category membership changes? Recent changes presents a firehose of information. Real-time, hourly, or even daily monitoring of recent changes is not what I need. Weekly or biweekly monitoring would require some kind of filter of the full recent changes, beyond what I already exclude. I think that might mean custom Javascript, which, except perhaps for tweaking, is beyond my paygrade. DCDuring TALK 15:52, 19 August 2015 (UTC)
Sysops are supposed to monitor Recent Changes in order to spot vandalism. It's a drag. SemperBlotto (talk) 15:57, 19 August 2015 (UTC)
I suppose if you limit Recent changes to the Category: namespace you'll only see changes there (including actual edits to categories, but there won't be many of those). What I want to know is, will things only show up if they've been added to the category manually (by typing [[Category:blah blah]] at the bottom, or will things also show up if they're categorized by means of a template? —Aɴɢʀ (talk) 16:27, 19 August 2015 (UTC)
The added functionality just provides an option to hide/show categorisation changes on Special:RecentChanges and Special:Watchlist. It's not possible to show only category membership changes. But I think, Aɴɢʀ's suggestion is a good workaround for that. Kai Nissen (WMDE) (talk) 17:38, 19 August 2015 (UTC)
Given what I can figure out about how this works, it doesn't matter. Think of it like this: whenever the software adds or removes an entry in a category, there's a "trigger" that gets run internally. This new feature merely adds a watchlist ping to one of the things that happens for that trigger. —CodeCat 16:34, 19 August 2015 (UTC)
Categorisation by templates is some kind of an edge case. To prevent the watchlist and recent changes from being flooded by the categorisation a widely used template, we limited the new feature to just log the template name and the number of pages embedding it. Kai Nissen (WMDE) (talk) 17:38, 19 August 2015 (UTC)
@Kai Nissen (WMDE): If I understood you correctly, that makes it utterly useless. I was hoping to be able to watch all the new entries in, for example, Category:Russian lemmas, into which pages are added almost exclusively by {{head}} and related templates. --WikiTiki89 18:21, 19 August 2015 (UTC)
It looks like it doesn't work that way. The recent changes are being flooded with bot edits to {{head}}. This wouldn't happen if templates were ignored as you say. —CodeCat 18:24, 19 August 2015 (UTC)
I think what Kai Nissen might have been talking about is what happens when you edit a template, rather than add a template to a page. --WikiTiki89 18:38, 19 August 2015 (UTC)
Exactly. Adding a template to a category results in "Template:Xyz and 1234 embedding pages added to category", while adding a categorising template to a page results in "New Page added to category". Kai Nissen (WMDE) (talk) 14:06, 20 August 2015 (UTC)
Can you just clarify one more time: Does that happen when adding the template itself to the category, or when making the template add pages it transcludes to the category? For example, what would the difference be between adding <noinclude>[[Category:Foo]]</noinclude> versus <includeonly>[[Category:Foo]]</includeonly> to a template (assuming there are already pages that transclude the template)? --WikiTiki89 14:14, 20 August 2015 (UTC)
The Category membership changes seem to only appear if one is explicitly watching the Category namespace, not by default (which is principal namespace only), based on my observations at MediaWiki, where it's been installed already. DCDuring TALK 16:37, 19 August 2015 (UTC)
Changes to category membership are always logged as a recent change. Once you add a category page to your watchlist, additions to and removals from that category will appear on your watchlist. The rollout on mediawiki.org is actually not a good example, since the edit (and especially categorisation) frequency is not that high. Kai Nissen (WMDE) (talk) 17:38, 19 August 2015 (UTC)
I think there are some quirks with bots and this change. DTLHS (talk) 18:13, 19 August 2015 (UTC)
It looks like the flood of bot edits has stopped, and the link to show/hide category changes is also gone. —CodeCat 18:19, 20 August 2015 (UTC)
And press "Show bots" does not make them show up. Strange... --WikiTiki89 18:23, 20 August 2015 (UTC)
I'm thinking it has been uninstalled altogether? —CodeCat 18:24, 20 August 2015 (UTC)
No, non-bot category changes are showing up. --WikiTiki89 18:26, 20 August 2015 (UTC)
Wait no, the bot ones are showing up. Maybe all the bots just happened to stop running. --WikiTiki89 18:27, 20 August 2015 (UTC)
One of them is mine, so I think I'd know. :P What is showing up is the regular bot edits. The category changes caused by those edits are not showing up. —CodeCat 18:51, 20 August 2015 (UTC)
It looks like they stopped being recorded after 18:10 UTC. The ones before that are still showing up. --WikiTiki89 19:09, 20 August 2015 (UTC)
I haven't seen any changes, by bots or humans, in a couple hours, nor have I seen any option to turn category-watching on or off. - -sche (discuss) 01:43, 21 August 2015 (UTC)

Dear all, as you have already noticed, this feature change got reverted on August 20th. After it got deployed to the production systems it became apparent that under some special circumstances it may cause privacy issues: In the case of some types of templates entries in the watchlist and recent changes were showing the IP address instead of the user name. It's good that this came to our attention as the developer team can start to work on a solution for this now. Thanks to all who already tested the new feature and gave valuable comments. It is a big help for the next version of this feature. Kai Nissen (WMDE) (talk) 09:56, 21 August 2015 (UTC)


I already see several problems with this extension the way it currently works.

  1. The category changes should be hidden from Recent Changes by default. Not sure about Watchlist.
  2. The extension ignores bot flags. Bot edits should not be shown unless it's explicitly enabled in RC.
  3. Entry names with spaces show up with "_", for example in my recent edit to killer games.

Any more? —CodeCat 18:30, 19 August 2015 (UTC)

The diff link just gives you the useless empty diff of the category. There should be a link to the diff of the page itself. --WikiTiki89 18:47, 19 August 2015 (UTC)
I agree that it should be hidden from recent changes and the watchlist by default. Otherwise, it overwhelms them. - -sche (discuss) 00:53, 20 August 2015 (UTC)
Let's make requests on Phabricator to report those issues, so the devs can correct them. — Dakdada 09:09, 20 August 2015 (UTC)
See phab:T109700 for the bots bug. — Dakdada 12:28, 20 August 2015 (UTC)
See phab:T109633 for hiding the category changes by default. — Dakdada 16:43, 20 August 2015 (UTC)

There seems to be a bug with language code categories. I added Category:en:Perching birds to a couple of pages - if I watch Category:en:Perching birds nothing appears on my watchlist, but if I watch Category:Perching birds it does. You can see this at Special:RecentChangesLinked/Category:en:Perching birds. Keith the Koala (talk) 15:24, 20 August 2015 (UTC)

what is this crap in recent changesEdit

So I create a new entry like "perillartine" and the Recent Changes shows three items: 1. me creating this (good!), 2. me adding it to the English lemmas category (bad! I didn't explicitly do this, and nobody cares), 3. me adding it to the English nouns category (bad! I didn't explicitly do this, and nobody cares). Who enabled this, and why, and how can we stop it, and how can we stop that person enabling such things in future? Equinox 02:36, 20 August 2015 (UTC)

Can you not see the section directly above this one? DTLHS (talk) 02:49, 20 August 2015 (UTC)
Oops. I should at least have read the "news for editors" page. Apparently everybody loves this change so it's just me against the world again, as with every version of MS Office since 2003. HOORAY. Equinox 02:51, 20 August 2015 (UTC)
You're not alone. Most people who like it, like it for the watchlist part, but are also complaining about it being enabled by default in recent changes. --WikiTiki89 12:04, 20 August 2015 (UTC)
You can turn it off with the "hide categorization" link, but I don't know if it's possible to make that permanent. DTLHS (talk) 02:54, 20 August 2015 (UTC)
e: there's an option in preferences under recent changes to hide categorization of pages. DTLHS (talk) 02:56, 20 August 2015 (UTC)
This seems to have fixed it, and my preferences should stick with my login, so I think we're good. Thank you! Equinox 02:58, 20 August 2015 (UTC)
Perhaps it's a nice capability, but it doesn't do much for me normally. I use and like the lists that can be placed in a category that show the most recent or oldest additions to the category.
I do like being able to see when something's been added to an entry maintenance category. —Aɴɢʀ (talk) 12:45, 20 August 2015 (UTC)
Then put the entry maintenance categories on your watchlist. There is really no purpose for this in recent changes. --WikiTiki89 13:03, 20 August 2015 (UTC)
Oh, I have. I was talking about the function in general, not its presence specifically on Recent Changes, which I almost never look at anyway. —Aɴɢʀ (talk) 13:35, 20 August 2015 (UTC)

Editing the Wikipedia link created by {{etyl}}Edit

Is it possible to directly link to Tat language (Caucasus) by {{etyl|ttt|xyz}}, rather than to the disambiguation page? --Vahag (talk) 16:53, 20 August 2015 (UTC)

I'll create this feature. It is currently available for etymology-only languages, but not for normal languages. --WikiTiki89 17:29, 20 August 2015 (UTC)
Yes check.svg Done. You can now do this. --WikiTiki89 17:41, 20 August 2015 (UTC)
The same may need to be done with language families. --WikiTiki89 17:43, 20 August 2015 (UTC)
Thanks, it's working. While we're on the subject of Tat, can you figure out why in Google Chrome a custom font is applied to Tat when sc=Latn, as in ձմերուկ ‎(jmeruk)? The same happens to Kurdish in Latin (code ku). The custom font is Segoe UI. My default is Arial. It does not happen to other multi-script languages, such as Udi. --Vahag (talk) 22:18, 20 August 2015 (UTC)
I have no idea. I've investigated this before for another language (I think Chinese) and just did again now. Without the lang="ttt" CSS parameter, this does not happen. But I cannot find anywhere how these language codes are associated with any CSS properties. --WikiTiki89 22:31, 20 August 2015 (UTC)
Sometimes browsers try to outsmart us. Here's an example of Japanese text marked with Latn script: Watch my font!. —CodeCat 22:31, 20 August 2015 (UTC)

Miscapitalized labelsEdit

Discussion moved from WT:T:TODO.

Is it possible to find cases like this? - -sche (discuss) 19:26, 20 August 2015 (UTC)

Couldn't the module just automatically test if the lower cased version of the label exists? DTLHS (talk) 19:32, 20 August 2015 (UTC)
That would also work, especially if we decide that the chances of labels being distinguished by capitalization (e.g. "art" meaning 'art' but "Art" being a dialect of, say, Hindi) are small enough to be negligible, which they probably are. - -sche (discuss) 20:20, 20 August 2015 (UTC)

This is an alternative form of a red linkEdit

Whatever we're doing to catch plurals with a red link for singular, could we also do it for "forms of" nonexistent lemmata, like logopœia (which currently links to a redlink, logopoeia). - -sche (discuss) 04:06, 21 August 2015 (UTC)

Changing Module:form of would probably be the most effective for tracking such cases across many templates. —CodeCat 06:30, 21 August 2015 (UTC)

No-L2-L3 tag for simplified Chinese redirectsEdit

My recent creation of 官网 was flagged with a "no-L2-L3" tag. For simplified Chinese words, it is customary to simply redirect them to their traditional counterparts, so any edit that uses {{zh-see}} should be exempt from the tag. -- King of Hearts (talk) 23:40, 22 August 2015 (UTC)


Is it possible to adapt this template for Bokmål? Donnanz (talk) 20:01, 23 August 2015 (UTC)


Could {{ja-def}} be updated to use the .Jpan CSS class? (I'd give it a shot but the template is locked) —suzukaze (tc) 05:38, 24 August 2015 (UTC)
(Why is the script .Latn anyways? The language has been defined in the template as ja...) —suzukaze (tc) 03:20, 6 September 2015 (UTC)

Missing pluralsEdit

A few months ago, I tried to catch missing English plurlals, but failed. At Module talk:en-headword I explained what I wanted to do. Could someone make it work, please? --A230rjfowe (talk) 10:07, 25 August 2015 (UTC)

Module error cropping up in many placesEdit

The following appears in many places, including at the top of this page and in sections preceding this one: <red>Script error: The module returned a value. It is supposed to return an export table.</red> —This unsigned comment was added by DCDuring (talkcontribs) at 00:24, 26 August 2015 (UTC).

I checked recent changes of the Module namespace, but no severely important module has been modified. Is there something wrong with the server? @Developers: It would also be helpful if this error message mentioned the name of the module that failed to return an export table. --WikiTiki89 00:26, 26 August 2015 (UTC)
I discovered it using Template:temp. DCDuring TALK 00:30, 26 August 2015 (UTC)
I wonder if anyone will fess up. DCDuring TALK 00:31, 26 August 2015 (UTC)
The problem seems to be fixed, with no new changes in the Module namespace. It was clearly a server-side software bug. --WikiTiki89 00:34, 26 August 2015 (UTC)
There definitely needs to be more communication between the developers/deployers and us. --WikiTiki89 00:37, 26 August 2015 (UTC)
Now that the problem is fixed, could someone with a bot do null edits on the 11,000+ pages in Category:Pages with module errors? As far as I can tell, there doesn't seem to me any change in the number of these aside from the results of the null edits I've done by hand. I'm guessing that, since there are no actual edits involved, the system isn't going to do anything with these by itself. I cleared all the category pages and a few of the mainspace entries, but there's obviously no way I can clear 11,000+ more. I should add that, since null edits don't affect the content of the entries, the edit histories, nor recent changes, even an unauthorized bot can do this with no ill effects or repercussions. Thanks! Chuck Entz (talk) 03:24, 26 August 2015 (UTC)
I'll do it if no one else is already (in which case let me know so we don't conflict). --WikiTiki89 04:04, 26 August 2015 (UTC)
By the way, what is the rate limit? I keep reaching it. --WikiTiki89 04:07, 26 August 2015 (UTC)
In case anyone cares, I figured out that the rate limit is 30 edits per 60-second interval. --WikiTiki89 06:32, 26 August 2015 (UTC)
Are you sure about that rate limit? When using pywikibot, I set the delay to its mininmum (1 second) and I get about 1 edit per second. Benwing2 (talk) 06:57, 26 August 2015 (UTC)
You have an approved bot. I don't. --WikiTiki89 07:05, 26 August 2015 (UTC)
Ah. OK. Benwing2 (talk) 07:08, 26 August 2015 (UTC)
Yes check.svg Done --WikiTiki89 12:54, 26 August 2015 (UTC)
Thanks. DCDuring TALK 13:27, 26 August 2015 (UTC)

User contributions oddityEdit

If I try to list the "user contributions" for User:Werdna Yrneh Yarg I get a blank page, the little rotating symbol in the top left and no response from Wiktionary. For anyone else it works properly. Any ideas? (Google Chrome under Windows 10) SemperBlotto (talk) 20:10, 26 August 2015 (UTC)

Have you tried resetting your cache? — Ungoliant (falai) 20:12, 26 August 2015 (UTC)
Yes. SemperBlotto (talk) 23:34, 26 August 2015 (UTC)
It works for me on Firefox on Windows, so I don't know what's going on; sorry. :/ Almost all of his edits are to talk pages; his edits to the main namespace stopped on the 11th and consisted of edits to toad, baby, bad, adze, neap, tup and neap tide. - -sche (discuss) 23:57, 26 August 2015 (UTC)
It works OK with Microsoft Edge - strange. SemperBlotto (talk) 00:10, 27 August 2015 (UTC)
I have the same problem (Windows 7, 10 + Chrome). Interestingly, when "Number of edits to show by default:" in the preferences is set to 50 (default) the page loads fine.--Dixtosa (talk) 17:47, 27 August 2015 (UTC)
Okay for me in Chrome on Win7. Equinox 17:54, 27 August 2015 (UTC)

This edit makes his page unloadable, apparently its summary's format being alien to MediaWiki is the culprit. @User:Werdna Yrneh Yarg, how did you make this edit? --Dixtosa (talk) 18:08, 27 August 2015 (UTC)

So sorry - I have just returned from leaflet distribution on the Lizard, 20 minutes ago. I believe that the asterisk may be the culprit, which I have now removed, because it no more applies. Shall avoid using that symbol in future, if it might cause problems. Removed asterisk (that relates to nothing visible), as I have found this borrowed word in Cornish. More seriously though, this may have caused the problem - very sorry about this. I could not access this talk page at first! Sorry to waste your time; but should any problem re-occur, please advise me, so that I can sort it promptly. Am now suspicious that someone is using a cyber attack - I hope that I am wrong. Werdna Yrneh Yarg (talk) 21:07, 27 August 2015 (UTC)Andrew

Extend TweenkBot voting a bit?Edit

See Wiktionary:Votes#User:TweenkBot_for_bot_status. The expiration of this vote was listed as August 25. I think we need to extend this vote -- up till the nominal expiration it only got 2 votes, both support but one of which was a support-with-conditions, which Dan seems to be doing now for all votes of this sort. I just submitted a (late) support vote and it would be nice if we could get a couple more people to weigh in one way or another so the vote is meaningful. Benwing2 (talk) 08:47, 27 August 2015 (UTC)

Don't votes normally last for a month? Why are bot votes (or just this bot vote?) initially set up to be shorter, anyway? - -sche (discuss) 17:27, 27 August 2015 (UTC)
Because people are eager to use their bots. --WikiTiki89 19:28, 27 August 2015 (UTC)

bad language code?Edit

In ку́рва ‎(kúrva), the {{cx|vulgar|lang=bе}} produces the error message The language code "bе" is not valid. —Stephen (Talk) 11:16, 28 August 2015 (UTC)

The language code was using Cyrillic е instead of Latin e. —Aɴɢʀ (talk) 11:27, 28 August 2015 (UTC)
Ah. That used to happen a lot, but it’s been a while since I saw it. Thanks. —Stephen (Talk) 12:08, 28 August 2015 (UTC)

Character codes in HTMLEdit

  1. I have used "& #0134;" in a number of places to represent a "dagger" (†), it has stopped working but "& #8224;" still works - what has changed? Is it temporary?
  2. Is "& mdash;" (for —) a recommended syntax or should we just use the character?   — Saltmarshσυζήτηση-talk 06:58, 29 August 2015 (UTC)
    &#0134; is equivalent to U+0086, which is a C1 control code for "Start of selected area", so I'm surprised it ever worked for you, and not surprised that it doesn't work now. You can also write &dagger; to get † just as you can type &mdash; to get —, but AFAIK it's always preferable to just insert the character itself directly, as it makes it easier for other editors to read the contents of the edit box. —Aɴɢʀ (talk) 07:08, 29 August 2015 (UTC)
What Angr said. Both of these characters can be inserting from the edittools below the edit window, and for the mdash you can also type {{subst:mdash}}. - -sche (discuss) 07:19, 29 August 2015 (UTC)
In Appendix:Greek verbs and a few other places, "& #0135;" and "& #0135; & #0135;" are also (as one might expect) not reolving into characters; what symbol(s) were they intended to represent? - -sche (discuss) 07:55, 29 August 2015 (UTC)
I think the double dagger ‡. In Windows, if you type Alt+0134 (using the number pad, not the numbers running along the top of the keyboard) you get † and if you type Alt+0135 you get ‡. —Aɴɢʀ (talk) 09:19, 29 August 2015 (UTC)
Appendix:Greek verbs also had some instances of &#0151; for the em dash. There seem to be a lot of Greek verb pages with these codes on them. Could someone with a bot go through Wiktionary and change all instances of &#0134;, &#0135; and &#0151; to , , and respectively? I wonder if there are others (potential ones include &#0138; for Š, &#0140; for Œ, &#0142; for Ž &#0150; for – (en dash), &#0154 for š, &#0156 œ, 158 ž, and &#0159 for Ÿ). These are old ASCII codes; see http://www.ascii-code.com/ and scroll down to "The extended ASCII codes (character code 128-255)" for the ones causing the trouble: the ones with decimal codes 128–159 are not in sync with Unicode points. —Aɴɢʀ (talk) 09:45, 29 August 2015 (UTC)
I've replaced them everywhere I could find them as of the July 2015 database dump. - -sche (discuss) 04:28, 2 September 2015 (UTC)
Thanks, -sche. —Aɴɢʀ (talk) 08:57, 2 September 2015 (UTC)
Thank you all. Obviously a "system characteristic", when I looked at {{el-T-Vs}} a few hours ago &#0134; was still showing the † - but not when viewed in preview. Anyway I'll amend my ways and use the wanted character. And thanks to Aɴɢʀ for your editing.   — Saltmarshσυζήτηση-talk 10:19, 29 August 2015 (UTC)

Bug regarding the Pronunciation section of Chinese entriesEdit

I have a bug to report. Basically there is a bug with the pronunciation template that is commonly used in Chinese entries. See this link for an example that doesn't have the bug: https://en.wiktionary.org/wiki/%E8%89%AF#Pronunciation

At the top right corner of that template, there is a button which says "Expand". The bug is that this "Expand"-button gets moved down if the entry also has links to Wikipedia. See these two entries: https://en.wiktionary.org/wiki/%E6%B0%B4#Chinese


VulpesVulpes42 (talk) 11:22, 29 August 2015 (UTC)

Reconstructed and appendix-constructed languages in Template:t, etcEdit

Previous discussion: Wiktionary:Beer parlour/2015/January#Can we insert proto-words into translation tables?

I'd like uses of Template:t, and if feasible also main-namespace uses of Template:head, to fail (or at least categorize into a cleanup category) when the language code supplied has its type set to reconstructed or appendix-constructed. This would prevent people from adding Klingon (etc) terms to translations tables, as in revenge, head and new. - -sche (discuss) 18:08, 29 August 2015 (UTC)

Done. —CodeCat 19:35, 29 August 2015 (UTC)
Thanks. So far only a dozen entries have shown up in Category:Pages with module errors, like channel (had a proto-language in the translations) and deer (had a Sindarin translation). - -sche (discuss) 19:58, 29 August 2015 (UTC)

Oldest/newest entries not working for reconstructed languagesEdit

@Daniel Carrero See Category:Proto-Germanic language. The lists remain empty. —CodeCat 20:33, 29 August 2015 (UTC)

@CodeCat That is because the list of oldest/newest entries in {{poscatboiler}} has a namespace= parameter. Usually the value is 0, I've made a workaround so that the value is 114 in categories named like Category:English citations and Category:English citations of undefined terms. Apparently it needs some kind of additional workaround so that when the language code is reconstructed (type = "reconstructed" according to Module:languages/datax) the namespace should be 100 (appendix). --Daniel Carrero (talk) 20:46, 29 August 2015 (UTC)
We should really get mw:Extension:DynamicPageListEngine installed. It allows multiple namespaces, and is callable natively through Lua. —CodeCat 21:41, 29 August 2015 (UTC)
Sounds good. I'd support that. --Daniel Carrero (talk) 21:55, 29 August 2015 (UTC)
As would I. How do we go about getting it? —Μετάknowledgediscuss/deeds 17:44, 30 August 2015 (UTC)
@Metaknowledge Created Wiktionary:Votes/2015-09/Installing DynamicPageListEngine. --Daniel Carrero (talk) 19:16, 29 September 2015 (UTC)

Appendix:I am (ethnicity): BurmeseEdit

The pipes in the Burmese transliteration (<span lang="" class="tr">kywantauka.ba.mapa||</span>) are being interpreted as wikicode for a new table cell. —suzukaze (tc) 01:08, 30 August 2015 (UTC)

Since User:Angr made the edit to Module:my-translit that caused this, maybe he can suggest a solution. Chuck Entz (talk) 02:56, 30 August 2015 (UTC)
I replaced the pipes with HTML entities, but I'm not sure what's going on with the testcases since they don't have | and || as the expected output. DTLHS (talk) 03:16, 30 August 2015 (UTC)
Now they do. I forgot to change the testcases when I changed what ၊ and ။ transliterate as. They used to transliterate as comma and period respectively, which was problematic because we also use the period to transliterate one of the tones, so I changed it to | and ||. —Aɴɢʀ (talk) 06:30, 30 August 2015 (UTC)

Bot task: Category:Persian pluralsEdit

This is the lone holdout in the plurals categories, with 88 entries. The few I've checked were originally created a number of years ago with the {{fa-arab}} script template instead of a headword template, the words "Plural form of " hard-coded, followed by another {{fa-arab}} with the singular, and the category also hard-coded. MglovesfunBot and MewBot swapped out the script templates, but not with the templates needed to make these into standard entries. Here's a typical example (already fixed) of what comes after the noun header:


# Plural form of {{l|fa|مبل}}

[[Category:Persian plurals]]

  • That obviously needs to become:

{{head|fa|noun plural form}}

# {{plural of|مبل|lang=fa}}

without the hard-coded category. They'll still be lacking transliteration and possibly other language-specific attention- but at least they'll be visible in the right categories. Chuck Entz (talk) 06:00, 30 August 2015 (UTC)

It's finished, I did it manually. Enosh (talk) 16:41, 30 August 2015 (UTC)
If I had known you were going to do it by hand, I would have done some of it myself. Oh, well. Thanks! Chuck Entz (talk) 21:16, 30 August 2015 (UTC)

{{calque}} does not accept etymology-only language codesEdit

See безволие ‎(bezvolije). --Vahag (talk) 22:16, 30 August 2015 (UTC)

The problem is that the template tries to combine {{etyl}} and {{m}}, but etyl uses etymology-only languages, and m shouldn't. I think someone is going to have to split the lang parameter into one for etymologies and one for linking to terms. Either that, or it will have to know what the regular-language equivalent is for a given etymology-only language, as in "la" for "ML.","NL.", etc. Chuck Entz (talk) 01:43, 1 September 2015 (UTC)
That information is already available through the "parent" value in the etymology language. So if the template is modified to use that if it detects that the code is an etymology code, then it should work. —CodeCat 02:14, 1 September 2015 (UTC)

Creating templates that drop the last vowel of the verbEdit

In Ojibwe, 1st person and 2nd person conjugations drop the final vowel if the verb ends with a short vowel (a,i,o). The most basic form and the most common form in dictionaries is the 3rd person. Take a simple verb, like Giiwe: "S/he goes home", Nin-giiwe: "I go home". Wiisini: "eats" is a verb that ends in a short vowel. Thus: Wiisini: "s/he eats", Ni-wiisin "I eat". How do I create a template that drops the final short vowel without creating a unique template for each entry? Ojibwe doesn't have very many entries or users, so, I've been trying to add some uniformity and templates to it, despite my lack of lua skills. MacBobart (talk) 23:33, 31 August 2015 (UTC)

This would need Lua though. Templates can't do any text processing other than sticking bits together. I wonder if there is a need for a Lua module that exposes and translates the Lua string functions so that templates can use them? —CodeCat 23:40, 31 August 2015 (UTC)
I thought that's what the module that Kephir renamed to "Ugly hacks" was for. I suppose that, for someone who can code in Lua, it's bad programming practice to piecemeal one's code like that, rather than moving it all to a module, but for those who only know templates, a string-function toolkit can really come in handy. Chuck Entz (talk) 01:35, 1 September 2015 (UTC)
As CodeCat says, templates can't remove or replace anything within a parameter, so you have to divide the word into the invariant part and the part that changes. In other words, you have a parameter for the part of the stem before the vowel, and another for the vowel that gets dropped. Then it's just a matter of creating the logic that adds the vowel in some forms or nothing in the forms that are truncated. Chuck Entz (talk) 01:35, 1 September 2015 (UTC)

September 2015

Double RSS postsEdit


Hello. I, the only person who uses RSS, would like to mention that each WotD seems to get posted twice on your feed (https://en.wiktionary.org/w/api.php?action=featuredfeed&feed=wotd&feedformat=atom). While I'm here, this also seems to happen to the Commons PotD feed (https://commons.wikimedia.org/w/api.php?action=featuredfeed&feed=potd&feedformat=atom&language=en) but not the FWotD feed. Anyway, sorry if this is the wrong place for technical complaints or if saying 'feed' too many times has made you hungry. Thanks! EDIT: FWotD doing this too btw. —This comment was unsigned.

User:Yair rand/newentrywiz.js should not link to elements in a head wordEdit

This now redundant (or usually redundant) as headword modules automatically link to individual words of multi-word terms. Renard Migrant (talk) 15:01, 1 September 2015 (UTC)

Fixed. --Yair rand (talk) 21:14, 1 September 2015 (UTC)

Latvian adjectives with missing headersEdit

There are about 2300 Latvian adjective entries with two headword lines, for two homographs, where the second headword-line is not preceded by a header. It should be straightforward to find and fix all such entries: find adj}}\n\n\{\{head\|lv\|adjective form}} and replace }\n\n\{\{head\|lv\|a with }\n\n===Adjective===\n{{head|lv|a (some entries have been fixed since the last database dump; the initial }\n\n\{ weeds those out). (If needed I can supply an outdated list of ~2600 such entries, of which a few hundred have been fixed.) Can someone with a bot do this? - -sche (discuss) 04:56, 2 September 2015 (UTC)

You overcomplicated the matter. Regex is not needed - a simple replace from
adj}}\n\n\{{head|lv|adjective form}} to
adj}}\n\n\===Adjective===\n{{head|lv|adjective form}}
does the job...--Dixtosa (talk) 17:20, 2 September 2015 (UTC)

Module:la-noun, a certain section not showing up, and irrelevant editsEdit

I recently added a ‘verbal nouns’ row to Module:la-verb so that verbal nouns derived from Latin verbs would be shown in the verbs' conjugation tables. I noticed, however, that for certain verbs, such as abeo and indico (in both senses), it simply wasn't showing up when it should have been. I thought it might have been the result of sloppy coding, but I noticed that after I made minor edits to different sections of the section for Latin on each page, it started showing up (sure enough, if I went to edit each conjugation table and hit ‘Show preview’, it would show up as well). What's going on here? Esszet (talk) 21:21, 3 September 2015 (UTC)

It sometimes takes a little time for each page that uses the template to get updated. Another thing you can do is purge the page (if you have the clock enabled in Preferences > Gadgets > User interface gadgets, then you can purge the page by clicking the clock). --WikiTiki89 21:30, 3 September 2015 (UTC)

Automated program to clean up names of mineralsEdit

Equinox created a large number of entries for minerals in this format. Although obviously each entry could use human attention, all entries of this kind could use two easy fixes: change the link to -ite into {{suffix||ite|lang=en}} (so at least it categorises) and change the context label from "mineralogy" to "mineral" (which displays the same way but again, categorises the entry correctly). Would anyone be willing to run this? —Μετάknowledgediscuss/deeds 05:42, 4 September 2015 (UTC)

"Aliases" in HotCatEdit

Is it possible/desirable to modify HotCat to handle "aliases", without slowing it down too much? For instance, as discussed in WT:RFM#Category:Occupations, it would be useful if someone could type "Professions" and HotCat would know to convert that to "Occupations", or might at least add "Occupations" to the popup "autocomplete" list of categories you might be thinking of. It might also help with people not knowing whether to add "Category:American English" (if someone had to add it manually for some reason) or "US English" and things of that sort. - -sche (discuss) 03:35, 6 September 2015 (UTC)

Category:Eastern EnglishEdit

This category is strange, and is probably the result of a recent change in Lua. Any ideas how to fix it? --Zo3rWer (talk) 20:48, 7 September 2015 (UTC)

It's a result of "east" and "eastern" being defined in Module:labels. When and why did they creep in? They seem too broad to be useful. Are they useful, or can they be removed? If they're kept, the short-term solution seems to be to add "eastern Pennsylvania" etc as an alias to "Pennsylvania" and then change "lunch kettle" etc to use "eastern Pennsylvania" rather than "eastern|_|Pennsylvania", the way "NYC" is currently an alias of the "New York" state label (because I didn't think there were enough NYC-specific words to merit separating them from NY state's words).
CodeCat recently did something that created a whole bunch of these categories, things like Category:Javanese English when people referred to the Java programming language in {{context}}, and awkward constructions like Category:North Korean Korean instead of Category:North Korean. I'm reluctant to create categories for these oddities, so Special:WantedCategories is full of them. Chuck Entz (talk) 21:14, 7 September 2015 (UTC)
Why not create the silly things and make then subcategories of something? I intend to do something similar for all of the pages that are "wanted" more than 50 times. DCDuring TALK 22:26, 7 September 2015 (UTC)
  • No, they should certainly not be created but instead dealt with, especially situations like Category:North Korean Korean that are splitting categorisation messily. @CodeCatΜετάknowledgediscuss/deeds 04:29, 9 September 2015 (UTC)
    The problem is very general: top down creation of categories, often of excessive fineness and often using generic categories without regard to how individual topics, languages, PoSes, etc are subcategorized in the world outside of Wiktionary. It seems that we are doing violence to reality to reduce it to something that we can have an "elegant" (read simple) technical solution for. In light of the apparently insufficiently flexible architecture, I am skeptical that the problems will actually be solved rather than minimally patched. The silly categories could be created and subcategorized into some kind of category of shame, with sensibly named categories inserted manually. I suppose we have to appreciate that the straitjacket imposed by our systems doesn't cause even more silliness and awkward wording. DCDuring TALK 11:28, 9 September 2015 (UTC)

Category:English words suffixed with -o-Edit

This is a MewBot-created category, three of whose four entries are my fault. How should -o- be correctly indicated, so that the category can be deleted? @CodeCatΜετάknowledgediscuss/deeds 23:39, 7 September 2015 (UTC)

Presumably the same as the other categories in Category:English words by infix. DTLHS (talk) 00:09, 8 September 2015 (UTC)
@DTLHS: But it's an interfix, not an infix, and we wouldn't want it to behave like one either. —Μετάknowledgediscuss/deeds 04:28, 9 September 2015 (UTC)
@Metaknowledge: See Category:English words by interfix then. --WikiTiki89 14:42, 9 September 2015 (UTC)

Quick Accel questionEdit

Typical situation: there is an {l} wrapped in an #if to display an mdash if there is nothing to put in the {l}, otherwise display the term in the {l}.

Is it better to wrap the WT:ACCEL spans specifically around the {l} (inside the #if) or it doesn't matter and they can be wrapped around the whole thing? (I kind of suppose there is absolutely no diff but wanted to check.) Neitrāls vārds (talk) 12:02, 8 September 2015 (UTC)

Just use {{head}}, it does all these things automatically. —CodeCat 15:26, 13 September 2015 (UTC)
CodeCat, I forgot to mention that this was in the context of inflection tables as opposed to headword line (and my orig. question can probably be disregarded altogether as there's probably zero difference.)
Also, can you add this to that page: User talk:Conrad.Irwin/creationrules.js#Please add (Moksha, Erzya, Latvian (nouns)? Or perhaps simply unprotect it, as I don't see how that page could have any appeal to vandals. Neitrāls vārds (talk) 22:55, 16 September 2015 (UTC)
For JavaScript, the protection is built in to the software and can't be taken off. That's a good thing too, because running scripts on a user's computer has a HUGE appeal to vandals. —CodeCat 23:31, 16 September 2015 (UTC)

Min Nan "o͘"Edit

In POJ there is a letter that looks like an o with a dot at the upper right. Wikipedia uses o + combining character "U+0358 combining dot above right" (o͘), but on Wiktionary it is a mixture of either the above or o + "U+00B7 middle dot" (o·). This leads to issues like 圖#Pronunciation (tô͘; combining dot above right) not properly linking to tô· (middle dot), and red links that point to o·-bá-sáng (middle dot) and not o͘-bá-sáng (combining dot above right). Is there a way to fix this with a bot? —suzukaze (tc) 04:23, 12 September 2015 (UTC)

Well theoretically that set of characters could appear in other valid contexts, so a bot probably wouldn't work. I can get you a list of pages and page titles where it occurs though. DTLHS (talk) 04:35, 12 September 2015 (UTC)
@suzukaze-c User:DTLHS/cleanup/min nan o DTLHS (talk) 05:00, 12 September 2015 (UTC)
Cool. I forgot to mention that there can also be tone markings on the o͘ (ô͘, ó͘, ò͘, ō͘). I suppose they'd become "ô·, ó·, ò·, ō·". —suzukaze (tc) 05:09, 12 September 2015 (UTC)
Alright, I can look for all of them. DTLHS (talk) 05:19, 12 September 2015 (UTC)
As DTLHS says, a bot might not work (perhaps if it were restricted to only editing spans tagged as Min Nan?); however, you could go through the list with AWB — that would probably be faster than going through it by hand. - -sche (discuss) 05:44, 12 September 2015 (UTC)
i've never been able to get awb to work on this computersuzukaze (tc) 05:59, 12 September 2015 (UTC)
The remaining instances relevant to Min Nan are mostly in the links to other-language Wiktionaries; should the links be removed? —suzukaze (tc) 02:54, 14 September 2015 (UTC)

Category tree bugEdit

It looks like {{poscatboiler}} is not correctly subcategorizing by-language meta-categories. E.g. Category:One-letter words by language is turning up under Category:Terms by lexical property by language, and not under Category:Character counts by language; or Category:Terms derived from DC Comics by language is turning up under Category:Terms by etymology by language, and not Category:Terms derived from fiction by language. --Tropylium (talk) 15:16, 13 September 2015 (UTC)

I don't see Category:One-letter words by language in Category:Terms by lexical property by language. —CodeCat 15:25, 13 September 2015 (UTC)
It wouldn't, because the change in the output of {{poscatboiler}} wouldn't propagate through to the categories without an edit. I just did a null edit on Category:One-letter words by language, and you should be able to see it in Category:Terms by lexical property by language now. This seems to be a case of all the intermediary links in the chain being skipped for umbrella categories, but not the language-specific ones. Chuck Entz (talk) 15:56, 13 September 2015 (UTC)
I still don't see it, no matter how many null edits I do. —CodeCat 16:36, 13 September 2015 (UTC)
I figured out where the disconnect is: the category in question is actually Category:Terms by lexical property subcategories by language. The confusion lies in the fact that the breadcrumbs are completely different for the umbrella categories as opposed to the language-specific ones, so it's easy to overlook that the one similar-looking part is different, too. If this restructuring was intentional, it wasn't handled very well. Everyone is used to the umbrella categories and the language-specific categories having exactly the same tree structure, so the obvious conclusion is that this is a bug. Chuck Entz (talk) 20:42, 13 September 2015 (UTC)
The "by language" categories never had umbrella categories, they always went into their own lumped-together category. Formerly, this category was decided by the specific category template being used, but then we merged them so it's just part of the data modules. The only exception has been the topical category tree, where the umbrella categories follow a structure parallel to the language-specific ones. —CodeCat 21:02, 13 September 2015 (UTC)

Thorndyke and othersEdit

A male surname? I think a template or module needs editing. SemperBlotto (talk) 15:40, 15 September 2015 (UTC)

Nope, just Wonderfolly. --Zo3rWer (talk) 15:45, 15 September 2015 (UTC)
By the way, in Czech, there are male and female surnames. --Zo3rWer (talk) 15:47, 15 September 2015 (UTC)
Not really, if it's anything like Russian: the surnames themselves are genderless, but there are male and female forms that match the gender of the person named (a man would be called by the masculine form, and a female would be called by the feminine form). It's very much like adjective gender in any number of languages: you say sabio for a man, sabia for a woman, and loco for WF... Chuck Entz (talk) 01:54, 16 September 2015 (UTC)
It seems the Wikipedia article Grammatical gender in Spanish forgot to mention the WF gender. Maybe it's too advanced a topic for an online encyclopedia. But back to the topic (or rather, the previous digression): Some people, even for Russian, consider male and female surnames to be separate lemmas, but I cannot agree with this view (after all, a brother and a sister really have the same last name). --WikiTiki89 02:00, 16 September 2015 (UTC)
I'd consider the feminine forms to be inflected forms of the masculine, just as with adjectives. —Aɴɢʀ (talk) 13:39, 16 September 2015 (UTC)
I seem to remember, when researching my Viking roots, that I have seen Norwegian surnames ending in -dottor as well as in -son. Would that be similar to the Russian usage? SemperBlotto (talk) 13:55, 16 September 2015 (UTC)
In some ways. In Russian the distinction is grammatical (for example, a surname ending in a consonant for the masculine nominative will usually have an -а added for the feminine nominative), but the Norwegian distinction of -son and -dottir is semantic, since these suffixes literally mean "son" and "daughter". The best way for you to understand the Russian system is with plurals, since English also sometimes pluralizes surnames (the Joneses) and in Russian, plural is just another gender/number class that also applies to surnames; thus, the same way you would use the plural form of a surname to refer to a family, you would use the feminine form of a surname to refer to a female member of the family. --WikiTiki89 14:13, 16 September 2015 (UTC)

template R:L&SEdit

Could someone please explain how this template works and show me where its parameters are explained? It seems one can but doesn't need to add the lemma e.g. R:L&S|pŏpŭlus. And why does it not work with that word? It should link to http://www.perseus.tufts.edu/hopper/text?doc=Perseus%3Atext%3A1999.04.0059%3Aentry%3Dpopulus1 --Espoo (talk) 20:10, 16 September 2015 (UTC)

  • You just need to look at the template's documentation. The first parameter is displayed, the second is the page name on Lewis & Short. I have fixed the usages on populus. SemperBlotto (talk) 20:22, 16 September 2015 (UTC)
Thanks for fixing that. My naive expression was in fact asking where the documentation is, but i found it now. It'd be nice if your fix could be used to better explain what the template's documentation unsuccessfully tries, and instead now uses the unhelpful and confusing expression "search term". In fact i wouldn't have understood your explanation either without the example since i would never have guessed that the lemma's page name on L&S is populus1.
This is probably the wrong place to ask for help with editing, but could you fix my edit "rare" as a label for the rare variant "adagio" on adagium? --Espoo (talk) 20:53, 16 September 2015 (UTC)

Module:pl-IPA (or maybe modules in general?)Edit

I'm trying to edit Module:pl-IPA, and I've noticed that when I try to do so, characters are occasionally entered one space behind that occupied by the cursor. I've never encountered this problem with other modules, but I highly doubt it would be specific to this one; is anyone else having this problem in this module or in any others? I'm using Safari on OS X 10.10.5. Esszet (talk) 17:20, 17 September 2015 (UTC)

Yes, I've noticed this when dealing with some Unicode punctuation. You can always do your editing offline and then copy and paste it in though. DTLHS (talk) 17:53, 17 September 2015 (UTC)
I did in spite of that what I wanted to do…can it be fixed? Esszet (talk) 22:33, 17 September 2015 (UTC)
@DTLHS Can it? Esszet (talk) 20:02, 20 September 2015 (UTC)


Shouldn't this template also put articles in a category, like Category:Requests for definition? Purplebackpack89 20:25, 17 September 2015 (UTC)

It does- maybe you have hidden categories turned off? Category:Italian entries needing definition DTLHS (talk) 22:44, 17 September 2015 (UTC)

Accelerated entry creation broken again?Edit

Some time around midday today Seattle time, green links stopped appearing. I was in the middle of creating several Japanese entries when this happened, and I was using the green links to accelerate the creation of hiragana and romaji entries.

I just checked the history at User:Conrad.Irwin/creationrules.js, which I think is the right code, and nothing has changed today. Can anyone help explain what's going on with our infrastructure, and (better yet) fix it? ‑‑ Eiríkr Útlendi │Tala við mig 00:01, 18 September 2015 (UTC)

CodeCat made an edit based on code written by Neitrāls vārds, which broke all accelerated creation links; I then reverted the edit. (There's a discussion about this at the Tea room currently.) Green links appear to be working again, but not for Japanese, AFAICT, which is odd. Perhaps they can sort this out. —Μετάknowledgediscuss/deeds 04:06, 18 September 2015 (UTC)
  • What's even weirder, they were working, then CodeCat's edit happened and they didn't work, then you reverted and they did work, then *something* happened on some other page entirely and they stopped working again. “WTF” is the general impression I'm left with. Our infrastructure seems entirely too fragile. ‑‑ Eiríkr Útlendi │Tala við mig 16:49, 18 September 2015 (UTC)
  • Ooookay... Accelerated entry creation for Japanese appears to be mostly working, except in a few oddball corner cases.
Could someone have a look at 什麼#Japanese, elucidate why accelerated entry creation isn't working for the hiragana and romaji spellings listed in the POS headline, and fix the issue? ‑‑ Eiríkr Útlendi │Tala við mig 21:53, 18 September 2015 (UTC)
I still can't figure this out. Once again, pinging @CodeCat, Neitrāls vārdsΜετάknowledgediscuss/deeds 22:28, 18 September 2015 (UTC)

English plurals and "These entry templates may help when adding words"Edit

The "Plural" button on Special:Search creates pages in the obsolete Category:English plurals category. —suzukaze (tc) 20:51, 20 September 2015 (UTC)

Fixed in diff. Enosh (talk) 12:24, 21 September 2015 (UTC)

edittools borkage: newlines now non-functionalEdit

As a New! IMPROVED! bug that just started appearing today, newlines included in my User:Eirikr/edittools page no longer render correctly in the edittools dialog, nor do they insert correctly when using the edittools dialog, appearing as the black-diamond � character instead.

What is going on here? ‑‑ Eiríkr Útlendi │Tala við mig 04:07, 21 September 2015 (UTC)

Category for pjt-conj-laEdit

Somebody ought to add a category to the template {{pjt-conj-la}}. --Lo Ximiendo (talk) 21:50, 21 September 2015 (UTC)

And {{pjt-conj-ra}} and {{pjt-conj-wa}}. --Lo Ximiendo (talk) 22:10, 21 September 2015 (UTC)
Added to Category:Pitjantjatjara verb inflection-table templates based of usage. Enosh (talk) 07:25, 22 September 2015 (UTC)

Broken Latin templatesEdit

Entries such as omnis, vulgaris, officinalis, and benghalensis are currently showing a whole lot of wikicode where the declension table is meant to be. It's hard to tell who or what broke it. Is anyone working on fixing this? @CodeCat? @JohnC5? —Pengo (talk) 02:50, 22 September 2015 (UTC)

Fixed now. Thanks. —Pengo (talk) 10:01, 22 September 2015 (UTC)

Also, is there a reason we have so much redundant wikicode in Module:la-adj/table? We ought to be keeping templates in the Template: namespace and calling frame:expandTemplate to use them (as I've done in Module:la-utilities). If la-adj/table was actually simplifying or generalizing something it might make sense, but it's just a text dump of a template into code, which offers little advantage and makes it much more difficult and error prone to edit. —Pengo (talk) 02:50, 22 September 2015 (UTC)

How would you handle the conditional inclusions? —CodeCat 19:50, 22 September 2015 (UTC)

Turn italic into normal in CategoryTreeEdit

The documentation of Template:table:chess pieces/en uses #CategoryTree to display a list of similar categories at the end of the page. (chess pieces, colors, suits, etc.) That list is in italic, can it be normal (as in, font-style: normal;)? I tried editing MediaWiki:Common.css but was unable to do it. --Daniel Carrero (talk) 19:15, 22 September 2015 (UTC)

@Daniel Carrero
body[class*="page-Template_table"] .CategoryTreeLabel {
should work —suzukaze (tc) 07:06, 28 September 2015 (UTC)
and possibly
body[class*="page-Template_table"] .CategoryTreeLabel:before {
for fake css bullets because the plain list is plain —suzukaze (tc) 07:09, 28 September 2015 (UTC)
It works when I use that CSS code on Special:MyPage/common.css, but it does not work when I use the same code on MediaWiki:Common.css (diff). Why is that? --Daniel Carrero (talk) 19:48, 28 September 2015 (UTC) Nevermind, working great! Thanks! --Daniel Carrero (talk) 21:53, 28 September 2015 (UTC)

extending sh-noun and other templates for multiple head parametersEdit

Could someone extend the sh-noun and other sh-* templates to work with multiple head parameters like it was done to sh-proper noun? It doesn't seem particularly complicated but I don't want to mess anything up as I'm not familiar with MediaWiki's templating syntax. Thanks. Fojr (talk) 15:15, 23 September 2015 (UTC)

Declension tables and multimedia objectsEdit

It seems that declension tables cannot be placed horizontally in line with multimedia objects (see here). I'm using Safari on a Mac; could this be the same bug as that which causes there to be a gap between the language header and the first sub-header underneath it when there's a FWOTD template along with a multimedia object between them in the source code (see here)? Esszet (talk) 14:56, 25 September 2015 (UTC)

Fixed: diff. As you can see, it is not declension tables in general, but specific templates that for some reason added clear: both to their CSS. You will likely have to fix this in other Slovak declension templates. --WikiTiki89 15:47, 25 September 2015 (UTC)


This template needs repair (or something).

See Category:List templates for use of {{poscatboiler|en|list templates}}. DCDuring TALK 04:53, 28 September 2015 (UTC)

What seems to be the problem? {{poscatboiler}} works well with Category:English list templates, which are templates with specific lists of words in English. Category:List templates uses the same name but is for a few random templates that do something else, including some code for formatting different types of lists. I suggest renaming it to something else or just dumping the contents into Category:Templates in case the category is not very useful by itself. --Daniel Carrero (talk) 05:20, 28 September 2015 (UTC)
Sorry, I meant Category:English list templates. The supplementary lists, often very useful, for oldest edited member of the category and newest additions to the category are empty in fact whereas there certainly should be members. DCDuring TALK 06:41, 28 September 2015 (UTC)
True, I see the problem now. There is an extension called DynamicPageListEngine that would help with that, which has been mentioned in a previous discussion about this problem. I created Wiktionary:Votes/2015-09/Installing DynamicPageListEngine to propose installing it; the vote explains how it would fix the problem, too. --Daniel Carrero (talk) 19:22, 29 September 2015 (UTC)

'Summary' namespaceEdit

I notice that we have a "Summary:" namespace with 12 pages in it: Summary:User talk:Geofferybard/Re: Cite policy, Summary:User talk:Internoob//ɵ/, Summary:User talk:Internoob/Checked again, Summary:User talk:Internoob/Dzogchen: When I switched it from Tibetan to English, I get the Union Jack flying over ancient Tibetan words...doesn't seem right, Summary:User talk:Internoob/Lobachevskyian, Summary:User talk:Internoob/assistance..., Summary:User talk:Internoob/message, vintage, Summary:User talk:Internoob/norveçli, Summary:User talk:Vpovilaitis/Nauja tema, Summary:User talk:Vpovilaitis/test, Summary:Wiktionary:LiquidThreads testing/more dots, Summary:Wiktionary:Sandbox/yes. The associated talk namespace has a single page in it: Summary talk:Wiktionary:LiquidThreads testing/more dots. Does this namespace serve a purpose that talk pages don't? WT:NS says it's for "Summaries of discussions in pages using LiquidThreads"; is it an integral part of LT? - -sche (discuss) 00:10, 30 September 2015 (UTC)

Yes. There's a "Summarize" link above every thread. Clicking on it opens an edit window for creating a Summary-namespace page. The text from that page then gets displayed after the words "Thread summary:" in the first message of the thread. It does look like nobody has used this feature in quite a while, though. Chuck Entz (talk) 01:35, 30 September 2015 (UTC)

No category in Russian for Orthodox Christianity?Edit

I was surprised to find out there's no such thing as Category:ru:Orthodox Christianity or similar. Could someone who understands the category system create it? Thanks! Benwing2 (talk) 05:48, 30 September 2015 (UTC)

How about Category:ru:Eastern Orthodoxy? You don't have to completely understand the category system to work with it: I went to Category:en:Roman Catholicism and went to its parent category to see what was there. I saw that there was Category:en:Eastern Orthodoxy, but there was no Category:ru:Eastern Orthodoxy, so I created Category:ru:Eastern Orthodoxy by putting {{topic cat}} in it.
If I had needed to add a category to the module, I would have clicked on the edit link in the text of Category:en:Roman Catholicism, which would take me to its data module, copypasted the section for Roman Catholicism at the right place in the alphabetical order, and replaced all the details to make the new section. The only way you can go wrong if you don't change the syntax is to specify a parent category that doesn't exist, which will give you a module error. Chuck Entz (talk) 12:39, 30 September 2015 (UTC)
Thank you. That works. I'll use your advice next time I need to create a category. Benwing2 (talk) 04:29, 1 October 2015 (UTC)

October 2015

Category:English timelines beginning in the 21th centuryEdit

Obvious problem is obvious; can anyone fix this? This, that and the other (talk) 02:41, 1 October 2015 (UTC)

So, the problem is "21th" where it should be "21st"? I'm going to fix that now. --Daniel Carrero (talk) 02:53, 1 October 2015 (UTC)
Thanks. FYI, you can move categories these days, rather than deleting and recreating. This, that and the other (talk) 04:15, 1 October 2015 (UTC)
Oh, I did not know that. Thanks for the info. --Daniel Carrero (talk) 04:27, 1 October 2015 (UTC)

Idea for a vandalism filterEdit

This would take a bit of messing about with regular expressions, but I've noticed quite a common habit of drive-by vandalism by simply changing the language header (e.g. "German" to "Russian"). Equinox 15:17, 1 October 2015 (UTC)

I suppose we could make this behavior generate a tag? L2-header-change or something. --Daniel Carrero (talk) 15:19, 1 October 2015 (UTC)
I would just prevent, because there's no way an anon would do that legitimately. If nothing else, there should be a filter for Ladin > Latin. —Μετάknowledgediscuss/deeds 17:40, 1 October 2015 (UTC)
Changing language headers is quite the sport these days (as it has been for a long time). Mostly it's changing "English" to anything that sounds foreign, with "Spanish" and "Chinese" being quite common. Then there are the people from different places in the Third World changing "English" to whatever is spoken in their country. The "Latin" > "Ladin" changes happen consistently every month or so, by IPs from all over the place- which makes it look more like ignorance than vandalism. Chuck Entz (talk) 01:53, 2 October 2015 (UTC)

Regional labels that apply to more than one languageEdit

At Talk:nae, someone brought up the fact that {{label|sco|Doric}} puts the word into Category:Doric Ancient Greek and provides a link to w:Doric Greek. It should, of course, put the word into Category:Doric Scots and provide a link to w:Doric dialect (Scotland). Another label that covers multiple languages is "Ulster": there is Ulster English, Ulster Irish, and Ulster Scots. In this case, the categories work correctly, but the Wikipedia links all point to w:Ulster, when it would more helpful for them to point to w:Ulster English, w:Ulster Irish, and w:Ulster Scots dialects. Unfortunately, I don't know how to edit Module:labels/data/regional to fix these labels, and any others that may be similar. (The "Munster" and "Connacht" labels also point to the Wikipedia articles on the provinces of Ireland rather than the articles on the dialects.) —Aɴɢʀ (talk) 17:13, 2 October 2015 (UTC)

This is why I recently made a strict separation between regional labels and dialects in the data modules. Of course, that doesn't actually help in this situation, but I was aware of that then. The current system doesn't allow you to give language-specific data for labels, which is a big shortcoming. But it should be pretty easy to add this in. —CodeCat 17:49, 2 October 2015 (UTC)
I just realise that making language-specific labels bring some new difficulties. Currently, aliases of labels are defined language-agnostically. If we introduce language-specific information, we may run into the situation where an alias only applies for a given language, but not for others. For example, if we have a label "Northern German" defined only for German, and a label "Northern Germany" as an alias of it, then something's going to break when someone uses it with "nds" as the language. It would also give a nonsensical result if it were the other way around: if "Northern Germany" were the main label, then "Northern German" could be used for languages other than German. —CodeCat 17:59, 2 October 2015 (UTC)
What about making the labels themselves language-specific, e.g. instead of {{label|grc|Doric}} and {{label|sco|Doric}} we have {{label|grc|grc-Doric}} and {{label|sco|sco-Doric}}, and likewise {{label|en|en-Ulster}}, {{label|ga|ga-Ulster}}, and {{label|sco|sco-Ulster}}? We'd only need to do that for ambiguous labels; those that are unambiguous could keep their current names. —Aɴɢʀ (talk) 18:43, 2 October 2015 (UTC)
That seems pretty pointless. We already provide the language in the template, why provide it a second time in the label? —CodeCat 18:52, 2 October 2015 (UTC)
Because "[t]he current system doesn't allow you to give language-specific data for labels, which is a big shortcoming." —Aɴɢʀ (talk) 22:59, 2 October 2015 (UTC)
Yes, but it's easy to allow language-specific data. The only difficulty is in the aliases. —CodeCat 23:44, 2 October 2015 (UTC)
Can we at least do something promptly about the fact that of the 19 terms currently in Category:Doric Ancient Greek, 11 are Scots? —Aɴɢʀ (talk) 07:18, 3 October 2015 (UTC)
To fix the immediate issue, create "Doric Scots" as a label. - -sche (discuss) 04:47, 4 October 2015 (UTC)
We do already have some labels with language-specific names, e.g. we have "Pomeranian Low German" (shortcut "Pomeranian LG") because I anticipated that bare "Pomerania"/"Pomeranian" could also apply to German, Polish, or the so-called Pomeranian languages. - -sche (discuss) 03:15, 3 October 2015 (UTC)

Categorization of homonymous suffixesEdit

In cases where a language has two homonymous derivational suffixes - is there anything we could do for distinguishing these in categorization? For example, Category:Finnish words suffixed with -lainen currently mixes two homonymous but etymologically, semantically and morphophonologically distinct suffixes. --Tropylium (talk) 01:15, 6 October 2015 (UTC)

You can use the id= parameter to distinguish them. —CodeCat 06:50, 6 October 2015 (UTC)

The pos=noun, pos=adjective, etc. parameter will distinguishing them in categorization. --Panda10 (talk) 12:32, 6 October 2015 (UTC)

That only works if they have different parts of speech, and even then it feels inferior to the id= method. —CodeCat 13:27, 6 October 2015 (UTC)
Read in another language