Wiktionary:Grease pit/2016/February

Some people have reported not receiving pings sent using {{ping}}. Anyone know why this happens? Someone once mentioned something about the ping needing to be in the same paragraph as the signature. Is this really true? (It seems quite suspect to me.) Benwing2 (talk) 07:01, 1 February 2016 (UTC)

User:Benwing2, And someone said it works when a new signature is added (or updated). This seems very logical to me.— This comment was unsigned.

I believe the ping is done as part of, or is triggered by, the expansion of the tildes. I've had good results with simply replacing my signature with the tildes when re-pinging, but I seem to remember someone saying it wouldn't work if you decreased the size of the post. Chuck Entz (talk) 13:34, 1 February 2016 (UTC)

I believe you may find this enlightening. —JohnC5 16:22, 1 February 2016 (UTC)

@Chuck Entz, JohnC5 Thanks! Benwing2 (talk) 01:08, 3 February 2016 (UTC)

@CodeCat I tried to put {{inh|rup|LL.|sclavus}} at the beginning of the etym section for shcljau and got an error "Lua error in Module:etymology at line 73: Late Latin is not an ancestor of Aromanian." Seems to me Late Latin IS an ancestor of Aromanian and other Romanian dialects. Is this fixable easily or is it a more basic bug in how {{inh}} deals with etymology-only languages? Benwing2 (talk) 07:14, 1 February 2016 (UTC)

@Benwing2 I have updated mod:languages/data3/r so that it should work now. —JohnC5 07:28, 1 February 2016 (UTC)

@JohnC5 Thanks! Benwing2 (talk) 07:46, 1 February 2016 (UTC)

A lot of entries in this category are due to someone using {{quote-newsgroup}} with both date= and year= set. A solution is to move the year into the date= parameter like . Would it be possible to fix these pages in that manner with a bot? - -sche (discuss) 22:26, 1 February 2016 (UTC)

@-sche: Yes. --kc_kennylau (talk) 10:52, 3 February 2016 (UTC)

At User talk:JohnC5#Greek Nasal Accent, there was recently a discussion about using {{head}} to populate categories like Category:English terms spelled with Æ automatically. I don't understand the technical bits, so here are the relevant parts of that discussion:

So, can this be done without too much hassle? — I.S.M.E.T.A. 23:45, 1 February 2016 (UTC)

It can be done, but I would propose a trial with just one language first, so that we can iron out any problems before it's done more widely. —CodeCat 00:02, 2 February 2016 (UTC)

Support. --Daniel Carrero (talk) 00:05, 2 February 2016 (UTC)

Support. An alternative is to have a bot go through the database dumps periodically and add categories, but ... this sounds like less work in the long run. - -sche (discuss) 02:40, 2 February 2016 (UTC)

Support, but I disagree with CodeCat and think that we should specify characters that we want categories for, rather than characters that we don't want categories for. Not all random characters that might happen to appear in a word are actually interesting. --WikiTiki89 19:48, 3 February 2016 (UTC)

@JohnC5, ObsequiousNewt, Wikitiki89, CodeCat, Daniel Carrero, -sche: So, shall we start the trial? — I.S.M.E.T.A. 15:44, 12 February 2016 (UTC)

@JohnC5, ObsequiousNewt, Wikitiki89, CodeCat, Daniel Carrero, -sche, I'm so meta even this acronym Template and module in use. Do not edit. --kc_kennylau (talk) 15:49, 12 February 2016 (UTC) Oops, forgot to strike --kc_kennylau (talk) 03:41, 13 February 2016 (UTC)

@kc_kennylau: How else do you suggest we put this into practice? — I.S.M.E.T.A. 01:28, 13 February 2016 (UTC)

@I'm so meta even this acronym: I have no opinion for now. --kc_kennylau (talk) 03:41, 13 February 2016 (UTC)

@kc_kennylau: Should your striking of your comment be interpretted as a withdrawal of your objection? — I.S.M.E.T.A. 15:02, 13 February 2016 (UTC)

@I'm so meta even this acronym: It means it is no longer in use and you guys may edit. --kc_kennylau (talk) 15:06, 13 February 2016 (UTC)

@kc_kennylau: I see. Were you editing the template and/or module? That wasn't very clear. — I.S.M.E.T.A. 15:08, 13 February 2016 (UTC)

@I'm so meta even this acronym: https://en.wiktionary.orghttps://en.wiktionary.org/w/index.php?title=Module%3Aheadword&type=revision&diff=37235828&oldid=36926076 --kc_kennylau (talk) 15:11, 13 February 2016 (UTC)

@kc_kennylau: Thank you! :-)  — I.S.M.E.T.A. 15:57, 13 February 2016 (UTC)

Category:User art-top and Category:User tokipona should probably be merged if possible. -Xbony2 (talk) 21:18, 3 February 2016 (UTC)

Done. The second category was created by User:Babel AutoCreate, a script (not a user or bot) which mostly adds valid cats whenever people use them on their user-pages. In cases where it adds invalid cats, the solution is to delete them and protect the pages so they can only be recreated by an admin. - -sche (discuss) 03:14, 13 February 2016 (UTC)

Hi ! I just created heahflod, using the prefix template in the Etymology section, and I'm seeing what looks like an attempt by it to establish a template (it's asking for documentation tab). Please help Leasnam (talk) 00:10, 4 February 2016 (UTC)

I see other entries are doing it as well. Please see *grunnatjaną, which shows it when trying to employ the suffix template Leasnam (talk) 00:14, 4 February 2016 (UTC)

See the next section. --WikiTiki89 02:12, 4 February 2016 (UTC)

At an entry like cat or Zika virus (See Zika virus#External links.) I am getting some material relating to documentation that is not supposed to be included. Why? I could not find a change that caused it, though my module foo is worse than my template foo. DCDuring TALK 00:17, 4 February 2016 (UTC)

I see that the problem is much more general, affecting my view of even this page. DCDuring TALK 00:21, 4 February 2016 (UTC)

See image. Someone seems to have broken MediaWiki somehow so that <noinclude> is now just unformatted text. See image above. I encountered the same problem on commons: while trying to upload the image. Renard Migrant (talk) 00:23, 4 February 2016 (UTC)

Aagh, me too! The whole site is wobbly, first noticed at pitar. —Aryamanarora ^{(मुझसे बात करो)} 00:24, 4 February 2016 (UTC)

Gosh, this is annoying. Almost no page on Wiktionary is displaying normally right now. Andrew Sheedy (talk) 00:26, 4 February 2016 (UTC)

I wonder how long it will take for this to flush out of the system once fixed. I wonder whether the system will have to go down for to accelerate the flushing. DCDuring TALK 00:31, 4 February 2016 (UTC)

Even WT:Main Page is broken. —Aryamanarora ^{(मुझसे बात करो)} 00:32, 4 February 2016 (UTC)

https://gerrit.wikimedia.org/r/#/c/255258/; some information regarding unclosed <noinclude> tags. —Aryamanarora ^{(मुझसे बात करो)} 00:39, 4 February 2016 (UTC)

Hi folks, sorry about that. I tried to change the behavior of unclosed XML-like tags to no longer eat everything afterwards on a page, which tends to break things horribly for most tags. It's being reverted now. Although I don't understand why you chose to rely on this behavior for apparently every single template with documentation on the whole wiki… Matma Rex (talk) 00:42, 4 February 2016 (UTC)

Saves keystrokes, I suppose...we are a little lazy. —Aryamanarora ^{(मुझसे बात करो)} 00:44, 4 February 2016 (UTC)

The ironic thing is, our laziness is what compels us to make thing easier. --Romanophile ♞ (contributions) 00:46, 4 February 2016 (UTC)

two sides of the same coin, I suppose. Also, the bug is back for me on विषाणु (viṣāṇu) —Aryamanarora ^{(मुझसे बात करो)} 00:55, 4 February 2016 (UTC)

What's the alternative to noinclude for template documentation? Renard Migrant (talk) 00:48, 4 February 2016 (UTC)

noinclude with a closing tag! — Ungoliant ^(falai) 00:51, 4 February 2016 (UTC)

I don't think anyone "relied" on this on purpose. It was probably an oversight in one template and then just got copy&pasted into other ones, as things go in MediaWiki land. Jberkel (talk) 00:59, 4 February 2016 (UTC)

I did it on purpose a lot. One thing I have learned is that people will use whatever works, and to stop them, you have to make it stop working. —CodeCat 01:09, 4 February 2016 (UTC)

Oh well then. Regardless of the reasons, it's definitely not safe for the parsing code to assume well-formed input. Probably a task for a bot to check & add the missing closing tags? Jberkel (talk) 01:42, 4 February 2016 (UTC)

To all the people I tried to get to stop doing this: I toldya so. --WikiTiki89 02:21, 4 February 2016 (UTC)

Can we make an edit filter for this (after existing unclosed tags have been closed)? DTLHS (talk) 02:27, 4 February 2016 (UTC)

For the less technically literate among us, what exactly is causing this problem? Is it a problem with templates themselves, or does it need fixing on individual pages? A lot of pages are still displaying like this for me, though oddly, some that were before now look normal. Andrew Sheedy (talk) 03:00, 4 February 2016 (UTC)

Imagine <noinclude></noinclude> as a box, where things go inside the box, such as <noinclude>HOW TO USE THE TEMPLATE:</noinclude>. Specifically, <noinclude></noinclude> means that the text inside should be shown when looking at a page (template, in this case) directly, but not when you include the page on another page (after all, no one needs to see "HOW TO USE THE TEMPLATE:" when trying to navigate an entry; that would be downright odd). For some reason a lot of the templates have boxes that aren't closed (<noinclude>). The box used to be automatically closed by the code that converts wikicode to HTML, but not anymore, so now all of the contents (template documentation) spill out onto the pages (because the software no longer sees the unclosed 'box' as "do not include this text when including the page on another page"). —suzukaze (t・c) 03:06, 4 February 2016 (UTC)

(e/c) The cause of the problem is that many templates used improper syntax (omitting the closing </noinclude> tag). This improper syntax used to work anyway, but the MediaWiki developers released an update that caused this syntax to stop working. The two potential solutions are: (a) to fix all our templates or (b) to roll back the software update. Option (a) is the only one we have control over, and is something we should do regardless of whether it is necessary. Option (b) is out of our control and probably not the best solution in the long run anyway. Essentially only templates need to be fixed, and some have already been fixed. --WikiTiki89 03:12, 4 February 2016 (UTC)--WikiTiki89 03:12, 4 February 2016 (UTC)

It looks like the update has been rolled back, but there are still lots of entries that won't be normal until the edit queue gets around to them or until someone does a null edit on each of them. I do recall an incident not that long ago when someone added categories or interwikis (I don't remember which) to the end of a template after the </noinclude>, and they ended up in every entry that used the template. That would be less likely to happen without a closing </noinclude>. Still, I don't like leaving loose ends like this hanging. Chuck Entz (talk) 03:50, 4 February 2016 (UTC)

Makes sense, suzukaze and Wikitiki, thanks. And while I figured out that null edits fixed the problem since the rollback, I wasn't sure if I was actually accomplishing anything, so I'm glad to learn that it is. Andrew Sheedy (talk) 03:58, 4 February 2016 (UTC)

The lingering effect seems to on search more than on the entries themselves. That is, the loaded entry pages generated by the search Chuck linked to above don't contain the nonsense. DCDuring TALK 04:01, 4 February 2016 (UTC)

I will go through the ones which are easy to fix, the ones which just need a closing noinclude at the end. I will also make a list of templates which have unbalanced noinclude and includeonly tags which will have to be cleaned up manually. - TheDaveRoss 02:59, 5 February 2016 (UTC)

I think that I got all of them, there are four vote templates which are messed up and someone should take a look at. Find them here. - TheDaveRoss 17:29, 6 February 2016 (UTC)

They seem fine to me. They are a bit confusing though, because they are meant to be substed and some of the tags are actually meant to be including themselves in the subst. --WikiTiki89 02:34, 7 February 2016 (UTC)

Perhaps we should add an edit filter to warn people that leaving an unclosed tag is a bad idea? - TheDaveRoss 14:19, 4 February 2016 (UTC)

If it could be done, .... Maybe we need special edit windows for template and module space that managed syntax and formatting. Too bad missing documentation couldn't be handled the same way. DCDuring TALK 14:56, 4 February 2016 (UTC)

I have never made an edit filter, but I think something along the lines of this would work for the two main offenders:

article_namespace == 10 &

!(

count("<noinclude>",new_wikitext) == count("</noinclude>",new_wikitext)

& count("<includeonly>",new_wikitext) == count("</includeonly>",new_wikitext)

)

Someone who actually knows what they are doing with these would have to look it over. - TheDaveRoss 13:19, 5 February 2016 (UTC)

On second thought, I just gave it a try and created Special:AbuseFilter/47. If those who know about such things would like to take a look that would be great. It seems to work, and gives a warning when there are unbalanced noinclude or includeonly tags. If regex is allowed in abuse filters I suppose it could work for any unbalanced tags, but that would get more complicated and more expensive. - TheDaveRoss 00:20, 7 February 2016 (UTC)

Of course regexes are supported there, but I don't think we should generalize this too much. --WikiTiki89 02:34, 7 February 2016 (UTC)

Looks good. Regexes are more complicated and processing intensive, I think this should be good enough for the majority of cases. Also checked, there's no whitespace allowed (e.g. </ noinclude>) and the tags are case sensitive (</Noinclude>), so we shouldn't need anything more complicated anyway. Jberkel (talk) 03:49, 7 February 2016 (UTC)

I wrote a tool to check for unclosed noinclude tags in an XML dump. Looks like most occurrences have already been fixed by @TheDaveRoss (thanks!). I fixed the remaining handful I came across except for {{documentation subpage}} which is locked. I'll rinse and repeat once a more recent dump gets published. Jberkel (talk) 02:53, 7 February 2016 (UTC)

Did I add the right </noinclude> to {{documentation subpage}}? DCDuring TALK 03:18, 7 February 2016 (UTC)

Yes. --WikiTiki89 03:36, 7 February 2016 (UTC)

Category:Englishproper nouns with unknown or uncertain plurals had two members per special wanted pages, now has three. It should have none. The category name is not hard-coded on the page. DCDuring TALK 03:03, 4 February 2016 (UTC)

Fixed. DTLHS (talk) 03:10, 4 February 2016 (UTC)

Thanks. DCDuring TALK 03:25, 4 February 2016 (UTC)

For some time (weeks?), clicking on the section header as for sheer on a page like WT:RFV (WT:RFV#sheer) would not take me to the section. Why? I assume this occurs because there is an obsolete section number. In contrast, the above named section link takes me to the right section. Does it happen to others? Is it something that can be corrected? What could I change in my behavior to reduce the annoyance? DCDuring TALK 13:04, 4 February 2016 (UTC)

Is there a template available for that? I can't find it. Donnanz (talk) 14:36, 4 February 2016 (UTC)

{{rfap}} Chuck Entz (talk) 14:52, 4 February 2016 (UTC)

Ah, cheers! Donnanz (talk) 14:59, 4 February 2016 (UTC)

I created multireflector recently. If you visit Category:English words prefixed with multi-, it's listed in both "Recent additions to the category" and in "Oldest pages ordered by last edit". Why should it appear in the latter? Does creating a page not count as an edit? If so, that stops this feature from being very helpful in finding long-untouched pages. Equinox ◑ 15:42, 4 February 2016 (UTC)

According to the specification the second list comprises of "pages sorted by date pages were last edit". I have removed "Oldest" part.

Interestingly the specs also adds

It should be noted, that lastedit really sorts by the last time the page was touched. In some cases this is not equivalent to the last edit (for example, this includes permission changes, creation or deletion of linked pages, and alteration of contained templates).

—Extension:DynamicPageList_(Wikimedia)

which does not sound very useful to me too. --Dixtosa (talk) 18:51, 4 February 2016 (UTC)

The sorting is reversed though, so the pages that have not been edited the longest show up first. —CodeCat 18:54, 4 February 2016 (UTC)

I purged the page and the problem seems to have resolved itself. So the issue was that it was not updated. --WikiTiki89 18:59, 4 February 2016 (UTC)

Romaji is no longer displayed on kana entries like フォークリフト and どうしょう. In addition, {{ja-noun|しょう の ふえ}} (簫の笛) should produce romaji "shō no fue", not "shōnofue".—suzukaze (t・c) 09:45, 5 February 2016 (UTC)

The last four edits by @kc_kennylau, CodeCat seem to have caused this... —suzukaze (t・c) 10:01, 5 February 2016 (UTC)

Now, I have a dim memory that somewhere on Wiktionary it was said that you can create templates using JS as well as Wikitext and Lua, but I can't find it anymore and it contradicts what the Wiki HQ says about Scribuntu. So I'm no longer that much convinced that it was about templates - or even more than a dream. Can JavaScript ever be used to edit mainspace pages? Korn (talk) 22:47, 7 February 2016 (UTC)

You can write whatever you want in js and then put it in either MediaWiki:Common.js or your common.js page. Is that what you meant? DTLHS (talk) 22:52, 7 February 2016 (UTC)

I meant whether I can {{#invoke:}} modules written in Javascript so I don't have to write e.g. conjugation tables in Wikitext. Korn (talk) 08:56, 8 February 2016 (UTC)

Nope. That's never been possible. --Yair rand (talk) 09:10, 8 February 2016 (UTC)

The template {{rfdef}} is malfunctioning. The documentation does not explain what is missing when I use the template on the entry for marvellously to request a definiton. Instead the template thinks a translation is needed and spouts something about Lua. In the event that the template is functioning correctly, then the documentation needs to be altered to match current template function. --EncycloPetey (talk) 01:41, 8 February 2016 (UTC)

These days it needs to be {{rfdef|lang=en}} (or whatever language). Equinox ◑ 01:46, 8 February 2016 (UTC)

I thought this was pretty clear: "Lua error in Module:utilities at line 97: Language code has not been specified. Please pass parameter 1 to the template." --kc_kennylau (talk) 01:53, 8 February 2016 (UTC)

I've brought the documentation up to date. —CodeCat 01:57, 8 February 2016 (UTC)

I think there should be more effort made by folks updating or converting templates/modules to ensure backwards compatibility. So many templates/modules are broken in historical revisions, and plenty of things are broken on live pages. In the case of the current topic it should handle a missing language parameter either by providing a default value or by gracefully failing (by categorizing the entry and providing the expected output). I support updating things to provide improved functionality, but it should be done properly. - TheDaveRoss 17:08, 8 February 2016 (UTC)

Is that going to happen as long as the folks "improving" the infrastructure mostly work on their own agendas? DCDuring TALK 17:37, 8 February 2016 (UTC)

It could be argued that throwing an error ensures that things get fixed quickly. Benwing2 (talk) 01:31, 9 February 2016 (UTC)

I am not saying that modules shouldn't throw errors, I am saying that if a template has been in use for years modifying it in such a way that historical revisions are all broken or worse live usage is broken, then that is a design failure. - TheDaveRoss 14:55, 9 February 2016 (UTC)

"Please pass parameter 1 to the template" sounds odd. I think it means please specify a first parameter. Renard Migrant (talk) 14:50, 9 February 2016 (UTC)

Re: TheDaveRoss I erroneously added notext=1 to {{inh}} ({{bor}} has notext and this doesn't) and it broke the entire template. In the pre-Lua days then simply the parameter would do nothing and in this specific case it would work perfectly. {{IPA}} has a similar problem. We seem to be getting less good at handling user input errors, not better. Renard Migrant (talk) 15:00, 9 February 2016 (UTC)

Is there a way to delete pages by bot? I ran a bot to create noun and verb forms for Russian lemmas and it created some erroneous forms due to errors in the declension templates on the lemma pages. I've gone through half the created forms and so far identified 14 such lemmas; this amounts to maybe 150 forms, too many to easily do by hand. Benwing2 (talk) 02:46, 8 February 2016 (UTC)

OK, I think I've figured it out. I assume however that I need to use an admin account, not a bot account ... is this correct? Or should I seek to have delete privileges added to my bot account (if there are such things)? Benwing2 (talk) 04:07, 8 February 2016 (UTC)

Nah, use your admin account and give it a flood flag. —Μετάknowledge^{discuss/deeds} 04:21, 8 February 2016 (UTC)

Hmm, what's a flood flag? Benwing2 (talk) 07:09, 8 February 2016 (UTC)

OK, figured it out I think, never mind. Benwing2 (talk) 07:18, 8 February 2016 (UTC)

FYI, if you ever need to delete a lot of pages more quickly than can be done using the normal method, but want slightly more control / opportunity for review before each deletion than you get with a bot, (a) AutoWikiBrowser can delete pages (if you log in with your admin account), and (b) quite a while ago Ruakh wrote some javascript, of which you can find a copy here, which can be modified (change "ri" to whatever you need, and the deletion reason to whatever is appropriate) so that when you land on the "delete page" page associated with entries starting with some string, it prefills the deletion summary and clicks delete; you can use that + Lupin / Navigation popups (which have a 'delete' button) to delete a lot of entries quickly but still "by hand". - -sche (discuss) 22:18, 8 February 2016 (UTC)

Thanks. I wish AWB worked better on a Mac. I've gotten it to work (sort of) through Wine, although it's a pain to install and it had some problems, if I remember ... I think it didn't respond properly to keyboard shortcuts. Benwing2 (talk) 01:29, 9 February 2016 (UTC)

Template:eo-spel is currently displaying two errors: at ehhoshangho chiujhaude it displays "ehhoŝanĝo ĉiuĵaude" instead of "eĥoŝanĝo ĉiuĵaŭde", and at chau it displays "ĉau" instead of "ĉaŭ". If someone more knowledgeable than me could fix this, it would be much appreciated. Pinging @Kc kennylau and @Robin van der Vliet, the two main editors of Module:eo-spel. —Mr. Granger (talk • contribs) 00:14, 9 February 2016 (UTC)

@Mr. Granger: Fixed. --kc_kennylau (talk) 03:00, 9 February 2016 (UTC)

Thanks! —Mr. Granger (talk • contribs) 03:07, 9 February 2016 (UTC)

Hi there,

I was trying to create a wiktionary entry for "MasculinitySoFragile" and was labeled as "probably spam." I'd like to report this as incorrect. Here is the code below

formatted the wiktionary entry wrong, despite it being a bad fit for inclusion. Thanks!]

— This unsigned comment was added by Anthoknees (talk • contribs) at 02:27, 9 February 2016 (UTC).

@Anthoknees Thanks for your post. Regarding the suggested entry, please see Wiktionary:Criteria for inclusion. Has this word been used in durably archived sources (generally meaning published books or Usenet posts) by at least three different people over a period of more than one year? If so, please provide evidence. If not, it does not meet Wiktionary's criteria for inclusion. —Mr. Granger (talk • contribs) 03:03, 9 February 2016 (UTC)

(edit conflict) This is a really bad entry for a number of reasons- so much so that I would delete it on sight if I saw it. First of all, please see our Criteria for inclusion- we never allow entries for something that someone just made up, unless it's clearly in widespread use independent from its creator and is likely to be around after a year. Please also be aware that this is a dictionary, not an encyclopedia, so all the biographical background info and interview material from the creator have to go (the fact that it's basically self-promotion just makes it worse). We strongly discourage links to personal web pages/blogs, not to mention commercial sites. Basically, all that's permissible would be the language header, a very short etymology header & section, the part of speech header followed by the {{head}} template (this is the only place it belongs), followed by a real definition, not a description of the term. Quotes in durably-archived sources showing the term being used are permissible, with the bare minimum of links. See Entry layout for details. By the way, "#" can't be used in an entry name, for technical reasons.

In other words, 90 percent of your text and almost all of the links (which are what triggered the edit filter) are inappropriate for an entry here- this isn't a blog, magazine, encyclopedia or link-farm. You also have a number of formatting problems, typos/misspellings and grammatical errors which would have to be fixed. Chuck Entz (talk) 03:13, 9 February 2016 (UTC)

• (too much edit conflict) @Anthoknees: I looked at the code of that abuse_filter and found out that the cause is your twitter links. --kc_kennylau (talk) 03:25, 9 February 2016 (UTC)

You're right, that's vandalism, not spam. Renard Migrant (talk) 14:49, 9 February 2016 (UTC)

Is there some configuration that we can change so that we can send thanks for edits made by IPs? — Ungoliant ^(falai) 14:41, 9 February 2016 (UTC)

I doubt that it is a configuration issue. I think the devs intentionally disallowed it, but I agree that there is no reason to prevent anons from being thanked. --WikiTiki89 16:13, 9 February 2016 (UTC)

This was requested on Meta a few months ago and is currently tracked as Phabricator:T63022. —Mr. Granger (talk • contribs) 21:15, 9 February 2016 (UTC)

{{calque}} doesn't seem able to deduce an etymology-only language's primary language; see móilín for an example. —Aɴɢʀ (talk) 11:17, 10 February 2016 (UTC)

@Angr: Because one would not link to an entry in an etymology-only language. --kc_kennylau (talk) 06:42, 11 February 2016 (UTC)

Etymology-only languages are variations on recognized languages. The modules should know that NL. is a variation of la, and link NL. terms to Latin entries. Chuck Entz (talk) 06:50, 11 February 2016 (UTC)

Exactly. If I write {{der|fr|VL.|*foobarius}}, it's interpreted as exactly the same as {{etyl|fr|VL.}} {{m|la|*foobarius}}, so {{der}} knows that the etymology-only language "VL." is to be matched with "la". {{inh}} knows that too. But {{calque}} doesn't. —Aɴɢʀ (talk) 07:31, 11 February 2016 (UTC)

T:calque perhaps needs a reform as well; the parameter names are too ugly. --kc_kennylau (talk) 08:20, 11 February 2016 (UTC)

First thing would be to rewrite it in Lua, then supporting etymology-only languages won't be hard. Benwing2 (talk) 08:40, 11 February 2016 (UTC)

In use. Please do not edit. --kc_kennylau (talk) 09:06, 11 February 2016 (UTC) Stroken 10:18, 11 February 2016 (UTC)

@Kc kennylau: Thanks for luacizing the template! But móilín is still getting a module error, just a different one now. —Aɴɢʀ (talk) 11:59, 11 February 2016 (UTC)

@Angr: Oh, it's the same, because the script cannot find the language for the code "NL.", which makes the value of "source" nil (nil = nothing). --kc_kennylau (talk) 12:33, 11 February 2016 (UTC)

This template was created so that descendants could be shown on different pages without duplication. It does its job well, but a major downside is that we have to create, maintain and watch a separate descendants page. I've been thinking about whether it's feasible to do this without the need for a separate page. Instead, the page for, say, Proto-Indo-European, would transclude the descendants from the Proto-Germanic page. That way, the descendants of each branch can be kept on their respective pages, while also being included on the ancestor page. Is this doable? —CodeCat 15:56, 10 February 2016 (UTC)

We frequently need (or ought) to add citations within the descendants tree structure. I have been hesitant in the past to add them to etymtrees. Can this be done safely? —JohnC5 16:06, 10 February 2016 (UTC)

@JohnC5: Try and tell me when you encounter a problem. --kc_kennylau (talk) 16:22, 10 February 2016 (UTC)

I've not encountered a problem as of yet. I am just curious about the behavior of a transcluded <ref> tag. —JohnC5 16:24, 10 February 2016 (UTC)

@CodeCat: I think it is better to put it on a separate page, because the display-text in the PIE page is not exactly same with the database page. Sure, it is still feasible to have different display-text on the same page. I am neutral in this vote. --kc_kennylau (talk) 18:33, 10 February 2016 (UTC) Stroken 18:41, 10 February 2016 (UTC)

Oops. I misunderstood. Then I still remain neutral. --kc_kennylau (talk) 18:41, 10 February 2016 (UTC)

@CodeCat: May I continue the etymtree project, since it seems like you're the only opponent. --kc_kennylau (talk) 03:36, 14 February 2016 (UTC)

What about implementing my proposal? Could you make that work? I'd much rather have descendant content on the same pages as the entries. Etymtree is a pain on the watchlist... —CodeCat 03:40, 14 February 2016 (UTC)

@CodeCat: I don't quite understand, since nothing is duplicated in the current format. --kc_kennylau (talk) 04:02, 14 February 2016 (UTC)

The issue I have with it is not the duplication (there is none as far as I can tell), but the need for creating a separate page to list all the descendants on. I would prefer to keep them in the entries. —CodeCat 04:05, 14 February 2016 (UTC)

@CodeCat: To quote your original comment:

he page for, say, Proto-Indo-European, would transclude the descendants from the Proto-Germanic page.

—CodeCat the Great

My response is there is no need to transclude those descendants, even in the current format. --kc_kennylau (talk) 04:57, 14 February 2016 (UTC)

Then why do we have etymtree? The point of it was to show all descendants, even those of sub-branches, without duplicating the information. I think we should not put the descendants on their own page. —CodeCat 18:46, 15 February 2016 (UTC)

Equinox ◑ 02:24, 12 February 2016 (UTC)

It was diff this edit by TheDaveRoss. DTLHS (talk) 02:27, 12 February 2016 (UTC)

Although there may be more of the same mistake in different templates- @TheDaveRoss DTLHS (talk) 02:32, 12 February 2016 (UTC)

Sorry about that, looks like a copy/paste error on my part. I went through the other template revisions I made and didn't see any further similar errors, of course that doesn't ensure that there are none. - TheDaveRoss 12:35, 12 February 2016 (UTC)

The audio boxes have gone black, and don't seem to be working. Hopefully it's just a temporary thing. Donnanz (talk) 14:16, 12 February 2016 (UTC)

For me, they've gone black, but they still work if you click on them. —Aɴɢʀ (talk) 17:08, 12 February 2016 (UTC)

The sound still isn't working for me, even allowing for the usual time delay. The menu comes up when clicked on though. Donnanz (talk) 17:29, 12 February 2016 (UTC)

They are finally back to normal. Donnanz (talk) 13:33, 5 March 2016 (UTC)

This edit did not trigger the filter. I tried two more edits, they also did not trigger the filter: this and this. --kc_kennylau (talk) 14:27, 12 February 2016 (UTC)

pinging @CodeCat. --kc_kennylau (talk) 14:28, 12 February 2016 (UTC)

I'm not very good with abuse filters. My last edit was meant to allow the prefix "doi" in links, which is not a language code but is an inter-project code similar to "wikipedia". You can revert it if you want, but please try to fix the problem I originally intended to fix. —CodeCat 17:11, 12 February 2016 (UTC)

@CodeCat: I'm not blaming you, but there has been no hits since 29 October 2015, which is coincidentally the date that you edited the filter. I'm not blaming you. I'm still finding out what caused the filter to be useless. --kc_kennylau (talk) 17:33, 12 February 2016 (UTC)

Well I have no idea! —CodeCat 17:35, 12 February 2016 (UTC)

@CodeCat: Found and fixed. --kc_kennylau (talk) 18:00, 12 February 2016 (UTC)

Thank you! —CodeCat 18:01, 12 February 2016 (UTC)

If the goal is to have the same functionality as the old filter and exclude "doi" as a prefix to match I think this is what you want:

(added_lines regex "\{2,3}):]+]") &
(good := "\{2,3}:" + rescape(article_text) + "\]\]";
!(added_lines regex good)) &
(alsogood := "\{2,3}):" + str_replace(rescape(article_text), " ", "_") + "\]\]";
!(added_lines regex alsogood))
& !(added_lines regex "\{2,3}):\{\{subst:(BASE)?PAGENAME}}\]\]")

The ((?!doi)pattern) does negative lookahead, so the {2,3} will not match doi. If that needs to be expanded, the doi can be changed to (?!(doi|w|ws|...) - TheDaveRoss 18:35, 12 February 2016 (UTC)

Or what Kc kennylau did. - TheDaveRoss 18:38, 12 February 2016 (UTC)

More than half of the Special:Tags had no documentation in the table, so I have written up the ones that I could. This leaves a handful that I don't understand. If they are yours (undocumented abuse filters, etc.) then please update the table with a brief description. Equinox ◑ 16:45, 12 February 2016 (UTC)

Could someone do a list of all the uses of {{a}} outside of pronunciation sections? A couple under ====Usage notes==== will likely be valid, but other than that it's occasionally misused for {{qualifier}} such as {{a|idiomatic}}. Some of the uses in pronunciation sections are technically incorrect too because of things like {{a|rare}} but they're a bit harder to find. Renard Migrant (talk) 17:55, 12 February 2016 (UTC)

@DTLHS. Renard Migrant (talk) 17:56, 12 February 2016 (UTC)

@Renard Migrant User:DTLHS/cleanup/a template outside section DTLHS (talk) 20:31, 12 February 2016 (UTC)

@Renard Migrant, DTLHS: Category:usage of a in wrong position, please report for false positives. --kc_kennylau (talk) 04:06, 13 February 2016 (UTC)

I'm not sure that is going to work, but ok. DTLHS (talk) 04:25, 13 February 2016 (UTC)

Rhymes:English/ɛðə(ɹ) is an obvious one. Wiktionary:Pronunciation doesn't count as a wrong usage either because it's discussing template usage in its text. Renard Migrant (talk) 18:32, 13 February 2016 (UTC)

@Renard Migrant: Thanks for the feedback. Adjusted the code so that only Main namespace is detected. --kc_kennylau (talk) 18:45, 13 February 2016 (UTC)

@Kc kennylau It's currently catching a lot of Pronunciation 1, Pronunciation 2, etc.. Could we fix that? —JohnC5 18:52, 15 February 2016 (UTC)

@JohnC5: I thought is prohibited. --kc_kennylau (talk) 13:27, 16 February 2016 (UTC)

• Could someone prepare a list of all the occasion where it is easier to type "qualifier" than "a". (I suspect that this might be easy.) DCDuring TALK 05:00, 13 February 2016 (UTC)

  ? If you don't want to type all the letters of "qualifier", you can type {{q}}. - -sche (discuss) 05:07, 13 February 2016 (UTC)

How much harder is it to type q than a? Renard Migrant (talk) 12:06, 13 February 2016 (UTC)

I couldn't imagine that there was a {{q}} to use, given that there were so many instances of {{qualifier}} in the wikitext I view. Is there a bot that changes {{q}} to {{qualifier}}? If so, why? DCDuring TALK 12:52, 13 February 2016 (UTC)

There are 80X as many uses of {{qualifier}} as {{q}}. DCDuring TALK 12:56, 13 February 2016 (UTC)

I think {{i}} is more common than {{q}} as a shortcut for {{qualifier}}, but yes, there are bots that convert both shortcuts to the long form, G-d knows why. —Aɴɢʀ (talk) 13:08, 13 February 2016 (UTC)

Perhaps Wiktionary needs the extra megabyte (0.25%) of size (assuming only one instance of {{qualifier}} per page)? {{i}} occurs on fewer than 500 pages. DCDuring TALK 13:36, 13 February 2016 (UTC)

The reason why bots should convert to the verbose form is that, unless one is already familiar with all of the "shortcut" templates (which even very high volume editors are not per this discussion) then the resulting wiki markup becomes incomprehensible. There are hundreds of possible templates which {{a}} might point to, laziness costs more time than it saves. {{accent}} would be a much better template name, {{dialect}} (or similar) better still. - TheDaveRoss 14:10, 13 February 2016 (UTC)

I have seen bots replace {{i}} with {{qualifier}} but I object to it. There's nothing wrong with renaming {{a}} to {{accent}} and making {{a}} a redirect; but I object to replacing shorter template names with longer ones. It makes life harder for editors, for one thing. It's not clear to me that {{qualifier}} is any more obvious to a neophyte than {{q}}. What I do think is a good idea is to standardize on one name, e.g. replace all {{i}} and {{italbrac}} with {{q}}. Benwing2 (talk) 03:17, 14 February 2016 (UTC)

I thought that we use {{lb}} for that kind of thing. I thought that {{a}} was for "attribute". Can we take a census of the types of uses of {{sense}}, {{label}} ({{lb}}), {{a}}, and {{qualifier}} ({{i}}, {{q}})? Do we have any single page that compares and contrasts these and, for that matter, {{gloss}} and similar others? DCDuring TALK 15:40, 13 February 2016 (UTC)

1. {{a}} = accent, used to label accents in pronunciation sections, goes before the pronunciation.
2. {{sense}} = sense, used to label senses in synonym sections, goes before the synonym.
3. {{lb}} = label, used to attach contexts like "metallurgy" or "physics" or "obsolete" to definitions; potentially adds the page to a category, goes before the definition.
4. {{gloss}} = gloss, used to gloss a definition by redefining it in different words, goes after the definition.
5. {{n-g}} = non-gloss definition, similar to a gloss but for non-gloss explanatory text that instead of a definition, often to explain the usage of grammatical particles.
6. {{i}} or {{q}} = miscellaneous italicized and parenthesized text.

Benwing2 (talk) 03:04, 14 February 2016 (UTC)

I appreciate the effort, but in what findable, permanent location will it reside? What path of links would a user follow to get to it? [[User: DCDuring

|DCDuring]] TALK 19:53, 15 February 2016 (UTC)

What do you think such a page should be named? IMO, each such template's documentation should point to the page describing them. Benwing2 (talk) 20:24, 15 February 2016 (UTC)

Relevance? The fact that {{q}} is a lot less used than {{qualifier}} unless I'm missing something doesn't make 'q' any harder to type. Renard Migrant (talk) 18:29, 13 February 2016 (UTC)

I think that Rukhabot (talk • contribs) broke the page. Now there are giant edit buttons floating on it. --Romanophile ♞ (contributions) 05:18, 13 February 2016 (UTC)

I think there was a problem somewhere else that got fixed. A null edit cleared it up (as far as I can tell). I always try a null edit before I report things like this. Chuck Entz (talk) 05:38, 13 February 2016 (UTC)

Weird. Thanks for fixing it, though. --Romanophile ♞ (contributions) 05:40, 13 February 2016 (UTC)

It's just a continuing, weird problem. I'm seeing it occasionally on the French Wiktionary too. Renard Migrant (talk) 18:28, 13 February 2016 (UTC)

See сытый--Dixtosa (talk) 12:20, 14 February 2016 (UTC)

this is what is should look like. --Giorgi Eufshi (talk) 07:22, 15 February 2016 (UTC)

Mglovesfunbot is renaming {{i}}, {{q}}, {{qual}} -> {{qualifier}}. I don't like this much; I think rather they should all be renamed to {{q}}. Now, if Mglovesfunbot can do this, what prevents me from running my own bot to rename them the way I like? And wouldn't this lead to a bot war? So I suggest that Mglovesfunbot shouldn't do this any more, and we should vote on which form to prefer. Benwing2 (talk) 17:15, 14 February 2016 (UTC)

Let Mglovesfunbot stop doing it. For one thing, the bot operator did not demonstrate consensus. For another thing, a recent vote for switching to certain short template names suggest a general preference for short names so the bot should not be switching things in the opposite direction. --Dan Polansky (talk) 20:28, 14 February 2016 (UTC)

I agree totally. It's still happening by the way, I had a load of these edits pop up on my watchlist the other day. DonnanZ (talk) 20:58, 3 November 2016 (UTC)

So don't watchlist them. Seems like an easier solution. Renard Migrant (talk) 23:13, 3 November 2016 (UTC)

I prefer to know what's going on with pages I have created. DonnanZ (talk) 10:55, 4 November 2016 (UTC)

@Renard Migrant Please stop doing these substitutions now. Bots are supposed to be making uncontroversial changes but this change is clearly controversial: I, Donnanz and Dan Polansky all object. You really should have stopped months ago, when this issue was first brought up. (If you don't stop, I may write a bot to undo these changes.) Benwing (talk) 07:26, 5 November 2016 (UTC)

It was uncontroversial at the time I started. Note that your bot would also be controversial because I would object to it, so that would be a violation of WT:BOT. Are you in favour of violating WT:BOT or against it? Seems you're against me doing it, but in favour of doing it yourself. Renard Migrant (talk) 11:16, 5 November 2016 (UTC)

You can be damn sure the next time your bot does something I will object to it. WT:BOT says 'I will make sure that the task is so innocuous that no one could possibly object.' Therefore any objection is enough to stop the bot. It doesn't matter why there is an objection. Renard Migrant (talk) 11:19, 5 November 2016 (UTC)

Cut the crap; this is a personal attack me pure and simple and has nothing to do with changing a redirect to a full form. If that were the case I'd have been blocked for about a million years already. Clearly none of you are moronic enough to think that somehow using the full name harms Wiktionary, so this is about me. So get on with it; what's the real problem here? Stop being cowards and lying about your true intentions. Renard Migrant (talk) 11:31, 5 November 2016 (UTC)

When I saw all these bot edits I was beginning to wonder whether {{q}} was deprecated, but I have now learnt that this is far from the case. No personal attack on RM intended though. I just want these edits to cease. DonnanZ (talk) 12:22, 5 November 2016 (UTC)

Just a small heads up that the modules that handle the data now supply a default value for these, so you no longer need to put in "type = regular", "scripts = {"None"}" or "family = qfa-und" anymore. Instead, you only specify a value if it's a real value, rather than an empty placeholder. The access modules (Module:languages, Module:families) fill these default values in automatically for now, so code that uses these works like it did before. —CodeCat 00:13, 15 February 2016 (UTC)

@CodeCat Could this change possibly have caused the error at wt:SCLIST? —JohnC5 00:41, 15 February 2016 (UTC)

Yes, but it's fixed. There are a handful of modules that process lots of data at once, and so they circumvent the normal access modules for speed. —CodeCat 00:44, 15 February 2016 (UTC)

Yeah, you fixed it just as I realized the issue. Thanks! —JohnC5 00:46, 15 February 2016 (UTC)

For me, the entry for "Thai" in that table displays a flag where the code should be, the code where the name should be, etc, with all the fields shifted to the right by one, resulting in an extra column in the table that only has a value ("9") on that one row. (Perhaps this bug is connected to the feature that puts flags next to language headers.) - -sche (discuss) 01:22, 15 February 2016 (UTC)

Yeah, that would be the flag extension, not the module. —CodeCat 18:40, 15 February 2016 (UTC)

Apparently WT:LOL, WT:LOF and WT:DATACHECK didn't get the memo. Please take a look. Chuck Entz (talk) 03:24, 16 February 2016 (UTC)

@Iudexvivorum, Wyang

User:Hippietrail has added {{attention|th|Which classifiers, counters, measure words are used with this noun?}} to พิซซ่า (pít-sâa), which is a missing feature, not a problem with the entry.

IMO, {{attention}} shouldn't be added to entries requesting additional info or features but to entries having some problems.

Could someone please add a cls= parameter to {{th-noun}}, similar to {{vi-noun}} or {{zh-noun}}?

The classifiers for the term are แผ่น (pɛ̀n) or ถาด (tàat). --Anatoli T. ^{(обсудить}/^вклад) 01:44, 15 February 2016 (UTC)

Sure, added. Wyang (talk) 03:50, 15 February 2016 (UTC)

Thank you! --Anatoli T. ^{(обсудить}/^вклад) 04:33, 15 February 2016 (UTC)

{{audio}} is being oddly displaced and covering text on the previous line; see Module_talk:cmn-pron. I think that the issue has to do with class="mw-collapsed" (see ), but that doesn't solve any problems. —suzukaze (t・c) 06:01, 15 February 2016 (UTC)

Thanks for raising it here Suzukaze. Removed the use of {{audio}} until this issue is resolved. Wyang (talk) 02:14, 16 February 2016 (UTC)

It seems to be OK for Chrome and Microsoft Edge, but now it's even worse for Firefox: instead of being half-hidden, the IPA is fully obscured (the audio box covers the entire. darned. line. wtf.). I don't know whether to cry or laugh —suzukaze (t・c) 07:04, 16 February 2016 (UTC)

Did something recently happen to {{audio}}? It used to be grey, but suddenly it's gone mostly black, which makes it difficult to see the duration of the recording and the "Menu" link. — SMUconlaw (talk) 17:19, 15 February 2016 (UTC)

OK, it seems to be back to normal. Not sure what happened, though. — SMUconlaw (talk) 14:58, 28 February 2016 (UTC)

Now that the {{cite-}} and {{quote-}} templates have been sorted out, would it be possible for someone to run a bot and do the following?

• Replace uses of {{reference-book}} with either {{cite-book}} (if it is used within <ref> tags or in ==References== sections) or {{quote-book}} (if it is used elsewhere, mostly to provide quotations for definitions of lemmas). The following parameters may need changing:
  • If |origyear= and |year= are used together, |origyear= → |year= and |year= → |year_published=.
  • |origdate= → |date=.
  • |origmonth= → |month=.
  • id=ISBN → isbn=.
• Replace uses of {{cite wikipedia}} with {{quote-wikipedia}}.
• Replace uses of {{cite-usenet}} and {{quote-usenet}} with {{quote-newsgroup}}.
  • There appear to be some uses of {{cite-usenet}} with the parameter "|prefix=#" or "|prefix=#*". These should be changed to "#* {{quote-newsgroup|...".
• Replace uses of {{cite book}} (a redirect) with {{cite-book}}.
• Replace uses of {{cite journal}} (a redirect) with {{cite-journal}}.
• Replace uses of {{cite news}} (a redirect) with {{cite-journal}}.
• Replace uses of {{cite web}} (a redirect) with {{cite-web}}.
• Replace uses of {{quote-news}} (a redirect) with {{quote-journal}}.
• Replace uses of {{reference-journal}} (a redirect) with either {{cite-journal}} (if it is used within <ref> tags or in ==References== sections) or {{quote-journal}} (if it is used elsewhere).
• Replace uses of {{reference-news}} (a redirect) with either {{cite-journal}} (if it is used within <ref> tags or in ==References== sections) or {{quote-journal}} (if it is used elsewhere).
• Replace uses of {{reference-song}} (a redirect) with {{quote-song}}.
• Replace uses of {{reference-us-patent}} (a redirect) with {{quote-us-patent}}.
• Replace uses of {{reference-video}} (a redirect) with {{quote-video}}.

The deprecated templates can then be deleted. Thanks. — SMUconlaw (talk) 17:53, 15 February 2016 (UTC)

I might be able to do that, after my current bot run (adding pronunciation to Russian entries) finishes. Benwing2 (talk) 18:19, 15 February 2016 (UTC)

Great, thanks! — SMUconlaw (talk) 18:35, 15 February 2016 (UTC)

Can you cite some examples that use "|prefix=#" and "|prefix=#*"? And some examples of {{reference-book}} inside of ==Reference== sections, some inside of <ref>...</ref> tags, and some that are neither? This will help with testing. Benwing2 (talk) 20:26, 15 February 2016 (UTC)

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘ Sure, here you go:

• {{reference-book}}
  • Used in "References" sections: 八, それ, あれ, lake
  • Used within <ref> tags: absurd, billion, guinea pig, slay
• {{cite-usenet}}
  • Using |prefix=: Obama Phone, pussygirl

— SMUconlaw (talk) 12:18, 16 February 2016 (UTC)

People still use Usenet? Had no idea ... Benwing2 (talk) 17:30, 16 February 2016 (UTC)

Apparently so! — SMUconlaw (talk) 21:46, 16 February 2016 (UTC)

It's inherent in the name: Usenet. But in all seriousness, I can't tell you if people still use it, but we certainly still quote it. --WikiTiki89 22:43, 16 February 2016 (UTC)

@Smuconlaw Some questions:

1. There are many cases where origdate= and origmonth= are present, but blank. In those cases, I just remove them, is this safe? Sometimes e.g. origmonth= is present and blank, and month= is present and non-blank.

   If the parameters are blank, it is safe to remove them. — SMUconlaw (talk) 07:08, 17 February 2016 (UTC)

2. What about templates in pages outside of the main namespace? I think we should fix 'Citations:', 'Appendix:', 'Reconstruction:' and certain others, but we need to leave alone 'Template:', and I'm not sure about 'User:', 'Talk:', 'Template talk:', 'User talk:', 'Appendix talk:', 'Category talk:', 'Wiktionary:', 'Transwiki:', etc.

   Hmmm. I agree you should fix "Citations:", "Appendix:" and "Reconstruction:", and would also suggest "Talk:", "Appendix talk:", "Category talk:", "Template talk:", "User talk:", "Wiktionary:" and "Transwiki:". It's probably best to leave the "Template:" and "User:" pages alone first – I'll need to see how the deprecated templates are being used there. — SMUconlaw (talk) 07:08, 17 February 2016 (UTC)

3. If id= is present but doesn't begin with 'ISBN ', currently I just leave it alone. Should I go ahead and rename to isbn= without stripping anything? Some examples:

• id=4-00-301001-9
• id=SBN 416 27990 2
• id=ISSN 0716730510
• id={{ISSN|00062510}}
• id=10-{{ISBN|4-000-60067-2}} (should I strip '10-ISBN ' in this case?)
• id=ISBN-13 978-0-00-722391-6 (should I strip 'ISBN-13 ' in this case?)

etc.

Benwing2 (talk) 06:37, 17 February 2016 (UTC)

I suggest dealing with these situations as follows:

• id=4-00-301001-9 → isbn=4-00-301001-9
• id=SBN 416 27990 2 → isbn=0-416-27990-2 (the SBN was the predecessor of the ISBN, and an SBN can be converted to an ISBN by adding a "0" in front)
• id=ISSN 0716730510 → issn=0716730510 (I suspect, though, that this particular example is an error as ISSNs only have eight digits. It's probably supposed to be an ISBN.)
• id={{ISSN|00062510}} → issn=0006-2510
• id=10-{{ISBN|4-000-60067-2}} (should I strip '10-ISBN ' in this case?) → yes, convert this to isbn=4-000-60067-2
• id=ISBN-13 978-0-00-722391-6 (should I strip 'ISBN-13 ' in this case?) → yes, convert this to isbn=978-0-00-722391-6

— SMUconlaw (talk) 07:08, 17 February 2016 (UTC)

@Smuconlaw I did a small run of the bot. Can you check the results in Special:Contributions/WingerBot? It's the most recent 50 entries or so. Especially check the ones that substituted {{reference-book}} and {{cite-usenet}}/{{quote-usenet}}. Note that one of the {{cite-usenet}} prefix substitutions left the prefix slightly wrong; I tried to account for cases where prefix= and an actual prefix already exist, but didn't handle this case quite right. Benwing2 (talk) 18:21, 17 February 2016 (UTC)

Also, I handled some of the cases above with strange ID's, but a few remain, and need to be handled manually. Here's the list:

• Page 10 Template talk:reference-book: WARNING: Would replace id= -> isbn= but isbn= is already present: {{quote-book| date = 1997-12-18
  • I think this will have to be manually handled. — SMUconlaw (talk) 15:26, 18 February 2016 (UTC)
• Page 65 lake: WARNING: Would replace id= -> isbn= but id=SBN 416 27990 2\n doesn't begin with 'ISBN'
  • As mentioned above, "id=SBN 416 27990 2" can be changed to "isbn=0-416-27990-2" (i.e., add a "0" in front of the SBN). If it isn't feasible to add the hyphens, leave them as spaces. — SMUconlaw (talk) 15:26, 18 February 2016 (UTC)
• Page 263 secondary structure: WARNING: Would replace id= -> isbn= but id=ISSN 0716730510\n doesn't begin with 'ISBN '
  • Change this to "issn=...". — SMUconlaw (talk) 15:26, 18 February 2016 (UTC)
• Page 838 Appendix:Glossary of traumatology: WARNING: Would replace id= -> isbn= but id=\n doesn't begin with 'ISBN'
  • I think this will have to be manually handled. — SMUconlaw (talk) 15:26, 18 February 2016 (UTC)
• Page 919 Talk:lathe: WARNING: Would replace id= -> isbn= but id= Reprinted by Lindsay Publications, Inc., Bradley, IL, USA. →ISBN (paper); →ISBN (cloth) doesn't begin with 'ISBN'
• Page 919 Talk:lathe: WARNING: Would replace id= -> isbn= but id= (Available as a reprint from Lindsay Publications in paperback →ISBN and cloth →ISBN doesn't begin with 'ISBN '
  • I think these will have to be manually handled. Wow, editors do put a lot of strange things in |id=. — SMUconlaw (talk) 15:26, 18 February 2016 (UTC)
• Page 1884 primary structure: WARNING: Would replace id= -> isbn= but id=ISSN 0716730510\n doesn't begin with 'ISBN '
  • Change this to "issn=...". — SMUconlaw (talk) 15:26, 18 February 2016 (UTC)
• Page 2304 Appendix:English third-person singular pronouns: WARNING: Would replace id= -> isbn= but isbn= is already present: {{cite-book| date = 1997-12-18
  • I think this will have to be manually handled. — SMUconlaw (talk) 15:26, 18 February 2016 (UTC)

Benwing2 (talk) 18:26, 17 February 2016 (UTC)

@Smuconlaw I'm doing another bot run now so you might have to search down a bit to find the start of the relevant changes to check (choose "500" at the bottom and search for "Xenu", the most recent entry of the relevant changes). Benwing2 (talk) 01:41, 18 February 2016 (UTC)

Had a random look at the some of the changes, and here are some thoughts:

• Occurrences of {{cite-usenet}} (a redirect) should be changed to {{quote-newsgroup}}. Similarly, {{quote-usenet}} (a redirect) should be changed to {{quote-newsgroup}}. (See Citations:evil laugh, Citations:macrophobia.)
• In Citations:frozen cow juice, this edit ended up with "##*" instead of "#* because there was already a "#" there. Similarly, if there was already an "*" which leads to "*#*", that would need to be fixed.
• In Citations:LTRFTW and other pages which originally used {{cite-usenet}}, |text= needs to be changed to |passage=.
• I see some occurrences like this: "|pages=p. 88" (haartätäkki). If there is a "p.", can this be changed to "|page=88"? Similarly, if there is a "pp." this should be changed to "|pages=88".
• Also, in abirritant and ability, an editor has typed something like "pages=6". Is it possible to change this to "page=6" if there is only a single number, and to "pages=6–10" if there are two numbers separated by a hyphen or en dash?

Thanks! — SMUconlaw (talk) 16:12, 18 February 2016 (UTC)

@Smuconlaw Rather than add special-case code for SBN and ISSN, these should be handled manually; the above cases are the only ones. Also, it looks like there are other parameters that need to be changed, e.g. looking at the definition of {{cite-usenet}}, it has params |group=, |googleid=, |link= that don't exist in {{quote-newsgroup}}. Similarly, {{quote-usenet}} has numbered params 2,3,4,5,6. Could you go through the various templates and make a full list? Also, why are the defns of {{cite-usenet}} and {{quote-usenet}} defined in terms of {{quote-web}} rather than {{quote-newsgroup}}? Benwing2 (talk) 18:51, 18 February 2016 (UTC)

OK, I see that {{quote-newsgroup}} does support |group= and numbered params 2,3,4,5,6, but not |googleid= or |link=. These extra params should probably be documented. Benwing2 (talk) 21:14, 18 February 2016 (UTC)

@Smuconlaw Haven't heard from you in a couple of days, do you still want the templates moved? Benwing2 (talk) 18:21, 20 February 2016 (UTC)

Sorry, give me a couple of days. I am away from home attending a conference. — SMUconlaw (talk) 18:26, 20 February 2016 (UTC)

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘ Hi, @Benwing2.

• When you replace {{cite-usenet}} with {{quote-newsgroup}}, please do the following:
  • Replace "|group=ABC" and "|googleid=XYZ" with "|newsgroup=ABC" and "|url=http://groups.google.com/group/ABC/browse_thread/thread/XYZ".
  • Replace |link= with |url=.
• And when you replace {{quote-usenet}} with {{quote-newsgroup}}:
  • Replace "|group=ABC" with "|newsgroup=ABC".
  • Replace "|text=ABC" with "|passage=ABC".
  • This is trickier, but if "|monthday=February 27" and "|year=2016" exist, can these be converted into a date like this: "|date=February 27 2016"?
  • For the unnamed parameters, replace them as follows:
    • {{{1}}} → |date=
    • {{{2}}} → |author=
    • {{{3}}} → |title=
    • {{{4}}} → |newsgroup=
    • {{{5}}} → |url=
    • {{{6}}} → |passage=

Thanks! — SMUconlaw (talk) 14:45, 27 February 2016 (UTC)

Hi, @Benwing2, just wondering if you saw the above message? — SMUconlaw (talk) 03:26, 8 March 2016 (UTC)

Oops, missed it. I'll get to this soon. Do any of the other templates that you want renamed have parameters that need changing? I wouldn't be surprised if they did. Benwing2 (talk) 03:29, 8 March 2016 (UTC)

At the moment, I don't think so. If you can work on the changes I mentioned in my 27 February post and my original 15 February post, that would be great. Thanks. — SMUconlaw (talk) 22:14, 10 March 2016 (UTC)

@Smuconlaw I'm working on this now. I think instead of having me set |url=http://groups.google.com/group/ABC/browse_thread/thread/XYZ you should support googleid= directly. In general you don't want to require people to insert lots of boilerplate like this, and furthermore the URL used by google could potentially change. Benwing2 (talk) 22:24, 15 March 2016 (UTC)

@Smuconlaw I did another test run of the bot. It's the last 100 entries or so at Special:Contributions/WingerBot. If you could check over some of them, that would be great -- if they look OK I'll go ahead and do all the rest. Benwing2 (talk) 09:07, 16 March 2016 (UTC)

Thanks, they look fine! — SMUconlaw (talk) 14:58, 16 March 2016 (UTC)

OK, I did the whole run. I had to fix up some templates because of use e.g. of text= when only passage= was accepted, in places where you didn't tell me to correct text -> passage. I'm sure there are other such issues; be alert for them. In general it sounds like you might need to be a bit more careful in the future when converting templates like this to make sure everything gets properly converted or supported. Benwing2 (talk) 06:21, 17 March 2016 (UTC)

Thanks again! — SMUconlaw (talk) 13:11, 19 March 2016 (UTC)

Problems are cropping up at a again. I don't think it's my fault this time. —suzukaze (t・c) 07:00, 16 February 2016 (UTC)

It's been doing it for about a week. The errors start in the middle of the French section, but I doubt there's any significance to that. I've been comparing the execution times of the different language sections, but I don't see anything obvious:

I would expect English to be highest in execution time, but Portuguese was a surprise. It does seem to be just a matter of lots of content, rather than a technical problem, though.

I would guess that something introduced by the recent edits to the modules has slowed things down enough in each link so that the total execution time goes over the limit (10.027/10.000 seconds). Chuck Entz (talk) 08:39, 16 February 2016 (UTC)

You are probably right. In the page source we read

<!-- 
Transclusion expansion time report (%,ms,calls,template)
100.00% 11540.434      1 - -total
 69.66% 8038.546     16 - Template:a
  4.62%  533.057     32 - Template:lb
  4.12%  475.782    708 - Template:l-self
  3.36%  387.314    288 - Template:m
  3.13%  361.671      9 - Template:list_helper_2
  2.82%  325.346    170 - Template:head
  1.83%  211.522    219 - Template:l
  1.41%  162.270     91 - Template:IPA
  1.22%  140.647     85 - Template:etyl
-->

So Template:a has the highest expansion time, and it has been recently edited. --Giorgi Eufshi (talk) 09:09, 16 February 2016 (UTC)

Revert. DCDuring TALK 13:33, 16 February 2016 (UTC)

Support revert --kc_kennylau (talk) 13:44, 16 February 2016 (UTC)

You should probably do this checking using a bot, either live or on the dump, and then create a user page containing all the errors. Benwing2 (talk) 17:26, 16 February 2016 (UTC)

If they ever complete Phab:T114072, then we will get not only the ability to auto-detect the language in many templates but also set up this sort of checking. But until that point, I think bots will be necessary. —JohnC5 17:37, 16 February 2016 (UTC)

Shall I revert this then? —JohnC5 18:40, 16 February 2016 (UTC)

Kenny was the author, AFAICT. I don't trust myself with any technical task, nor should others trust me. So, if Kenny hasn't and you trust yourself, then go ahead. DCDuring TALK 19:43, 16 February 2016 (UTC)

But Kenny has attempted a repair, so let's see if it works. DCDuring TALK 19:45, 16 February 2016 (UTC)

That repair ensures that the checkusage function (the culprit here) only runs in the mainspace. This does not prevent the checkusage function from continuing to be extremely expensive. —JohnC5 20:01, 16 February 2016 (UTC)

Indeed. ] still shows gobs of module errors, most of the bottom of the page. Why are no other pages effected? DCDuring TALK 22:56, 16 February 2016 (UTC)

Other pages are not affected because ] is a huge page. Someone probably slowed down a widely transcluded template by a little, which caused huge problems at ]. --WikiTiki89 23:02, 16 February 2016 (UTC)

I've removed the check, fixing the issue. And yes, if we want to do this kind of checking, it should be done by bot, not in modules. I left the function there for reference, but it should not be used in actual invocations of {{a}}. --WikiTiki89 23:07, 16 February 2016 (UTC)

Thanks. I had looked at other large pages, like ], which had no such problem. DCDuring TALK 23:24, 16 February 2016 (UTC)

The water entry is actually not that big (11 language sections)- it just has a planet-sized translation section imbedded in it. Changing the subject: let's not overlook Module:labels, which has similar logic added by Kenny at about the same time. It's not nearly as much of a time-hog as Module:a, but if you look at the next template in the list above, Template:l-self uses less execution time than Template:lb in spite of being used 22 times as much. Chuck Entz (talk) 03:31, 17 February 2016 (UTC)

Do we have lists of the entries that are the top users of each template or use the largest number of templates, the top 10 entries in size, the top 10 L2 sections? Some such lists seem like they would be useful for testing performance issues. Such a set of lists could be taken from the dumps, which include a 250Mb file of "Wiki template inclusion link records". Someone more knowledgeable than I could probably think of more useful lists, also.

Wouldn't the useful ones supplement special pages like Special:MostTranscludedPages to help developers and users predict and check on the consequences of template and module changes? DCDuring TALK 13:57, 17 February 2016 (UTC)

I just noticed the logic to detect wrong placement in Module:labels: it's horribly inefficient and definitely the wrong place to perform these kind of checks. As a general rule, modules should only work on the data which is passed to them via parameters, not by fetching the whole content of the surrounding page. It's very brittle and likely to blow up any time. Can somebody please revert it? Jberkel (talk) 13:50, 24 February 2016 (UTC)

I agree that modules shouldn't parse the entire page in which they are transcluded, especially one such as this which is used all over the place. The section in question doesn't provide and front-end functionality (I think) it just categorizes errors. @Kc_kennylau can we just comment the section out for now? - TheDaveRoss 14:20, 24 February 2016 (UTC)

I've added a bit to WT:LUA that suggests previewing potentially time-consuming edits against a large page like ]. - -sche (discuss) 17:47, 25 February 2016 (UTC)

I was curious how many {{label}}s are actually incorrectly tagged and wrote a script to perform an offline analysis. The answer: not that many. Surprisingly there are only around 500 instances which means I could have overlooked something. Anyway, feel free to help clean up entries in "your" language. – Jberkel (talk) 02:36, 26 February 2016 (UTC)

So, I know that water wasn't malfunctioning before and that this may not be {{a}}'s fault, but water is now running out of execution time near the end. —JohnC5 22:16, 26 February 2016 (UTC)

It seems that even {{t-simple}} isn't fast enough, we should probably just flatten out the templates on that page. - TheDaveRoss 22:22, 26 February 2016 (UTC)

Here's the breakdown:

<!-- 
NewPP limit report
Parsed by mw1012
Cached time: 20160226195113
Cache expiry: 2592000
Dynamic content: false
CPU time usage: 14.910 seconds
Real time usage: 15.690 seconds
Preprocessor visited node count: 47243/1000000
Preprocessor generated node count: 0/1500000
Post‐expand include size: 476181/2097152 bytes
Template argument size: 38886/2097152 bytes
Highest expansion depth: 10/40
Expensive parser function count: 1/500
Lua time usage: 10.148/10.000 seconds
Lua memory usage: 39.09 MB/50 MB
-->

<!-- 
Transclusion expansion time report (%,ms,calls,template) 
100.00% 13063.157      1 - -total
 28.09% 3669.072   1539 - Template:t-simple
 22.15% 2892.920    688 - Template:t
 15.99% 2089.391    613 - Template:l
 11.01% 1437.839     25 - Template:label
  4.63%  605.436    167 - Template:t+
  1.74%  227.514     41 - Template:etyl
  1.67%  218.260     44 - Template:m
  1.33%  173.735     62 - Template:t-needed
  1.24%  161.955     13 - Template:IPA
-->

You'll notice that Template:label is 4th on that list, in spite of only 25 calls to t-simple's 1539 calls. Let's take the known outrageously inefficient code out of there before we try to squeeze anything more out of t-simple. Also, this is an intermittent problem: after a null edit, it was back to normal, with 5.4 seconds total script execution time on the preview before I saved. When I viewed the source, it was back up to 8.2 seconds:

<!-- 
NewPP limit report
Parsed by mw1004
Cached time: 20160227013323
Cache expiry: 2592000
Dynamic content: false
CPU time usage: 12.211 seconds
Real time usage: 12.630 seconds
Preprocessor visited node count: 47256/1000000
Preprocessor generated node count: 0/1500000
Post‐expand include size: 475799/2097152 bytes
Template argument size: 38886/2097152 bytes
Highest expansion depth: 10/40
Expensive parser function count: 1/500
Lua time usage: 8.201/10.000 seconds
Lua memory usage: 38.86 MB/50 MB
-->

<!-- 
Transclusion expansion time report (%,ms,calls,template)
100.00% 10530.839      1 - -total
 28.08% 2956.537   1539 - Template:t-simple
 21.90% 2306.068    688 - Template:t
 14.26% 1502.104    613 - Template:l
 11.52% 1212.923     25 - Template:label
  5.56%  585.753    167 - Template:t+
  1.88%  198.237     44 - Template:m
  1.84%  193.439     41 - Template:etyl
  1.18%  124.175     13 - Template:IPA
  1.16%  122.598     62 - Template:t-needed
-->

My theory is that background tasks slow things down from time to time, enough to go over the limit. Chuck Entz (talk) 01:44, 27 February 2016 (UTC)

• Can someone check colon#Taanslations. I think the part under header "The last part of the digestive system" is misplaced. Borovi4ok (talk) 09:56, 16 February 2016 (UTC)

I can't see anything. 'Part' is a bit vague, do you mean one of the templates? Renard Migrant (talk) 14:00, 16 February 2016 (UTC)

I think he means that the second translation table (last part of the digest system) under Etymology 1 (punctuation) belongs under a different Etymology section. DCDuring TALK 23:17, 16 February 2016 (UTC)

• Thanks @Borovi4ok. DCDuring TALK 23:17, 16 February 2016 (UTC)

See e.g. User:Visviva/NYT_20090426. The template works in mainspace but apparently not on Visviva's pages, though it was fine until recently. Equinox ◑ 10:53, 17 February 2016 (UTC)

Wrong question. Should be "who broke {{User:Visviva/quote-news-special}}?". It looks like a side-effect of the effort by @Smuconlaw to rework all of the quote templates, though I haven't tracked down the specifics. Chuck Entz (talk) 13:38, 17 February 2016 (UTC)

Yes, I noticed that. Had a quick look yesterday, but can't figure out what's wrong. Also, I don't know if it was already showing errors before any changes were made. Let me know if you can figure out what's wrong. — SMUconlaw (talk) 13:42, 17 February 2016 (UTC)

Judging by Special:WhatLinksHere/Template:start_year, this would seem to be part of a larger problem. I don't know enough about templates to figure it out. Chuck Entz (talk) 13:47, 17 February 2016 (UTC)

Had another look. There's no {{start year}}, so I don't know why there are pages linking to it. I've never worked on that (non-existent) template before. — SMUconlaw (talk) 13:51, 17 February 2016 (UTC)

OK, I figured out what happened. These pages use {{reference-journal}} which has now been renamed {{quote-journal/source}}, and there was a typo in the |start_year= parameter. I think the problem is now solved (you may need to do a page purge to see the change take effect). — SMUconlaw (talk) 14:14, 17 February 2016 (UTC)

Thank you! Glad to see I pinged the right person. I've done null edits on all the mainspace uses, which did clear them. There are still 569 pages left, but I don't have the patience, let alone the time, to deal with those. Chuck Entz (talk) 15:04, 17 February 2016 (UTC)

In the deletion reasons list, what happened to "intimidating behaviour/harassment"? I saw no vote about its removal. Equinox ◑ 22:58, 18 February 2016 (UTC)

Was that ever in the deletion reason list? It's a block reason. DTLHS (talk) 23:18, 18 February 2016 (UTC)

Huh, oh yeah, you're right. Equinox ◑ 23:20, 18 February 2016 (UTC)

I tried to add the "Alternative forms" section to as-yet-unknown and got the warning that the change was auto-identified as harmful and disallowed. This is some false-positive somewhere that needs to be corrected. 73.71.174.75 00:26, 20 February 2016 (UTC)

You added a faulty "===Alternative forms====", which has 3 equal signs in the front but 4 equal signs in the back. --kc_kennylau (talk) 04:10, 20 February 2016 (UTC)

That doesn't seem to me to be a severe enough mistake to warrant a "harmful change" warning. —Aɴɢʀ (talk) 09:03, 20 February 2016 (UTC)

Agreed. That's tag's job to detect that kind of edits.--Dixtosa (talk) 15:04, 20 February 2016 (UTC)

So now this gadget can be used with the languages with automatic transliterations superpowers. Here's the modified version. and this is the diff. @Yair rand. --Dixtosa (talk) 18:21, 20 February 2016 (UTC)

Merged. --Yair rand (talk) 11:53, 21 February 2016 (UTC)

See for example armee which has the plurals armees but the automatic module-generated plural is armes. I assume it's just a pure error and not intentional. Also specifying manually the plurals armees adds it to Category:Old French irregular nouns which of course, couldn't be further from the truth, it's just a case of manually fixing a module error. Hopefully that one's fixed, can anyone find and fix or find and tell me where the -ees to -es error is? Renard Migrant (talk) 18:08, 21 February 2016 (UTC)

@Renard Migrant This should be fixed now. Benwing2 (talk) 22:17, 21 February 2016 (UTC)

Looks good. Renard Migrant (talk) 12:37, 23 February 2016 (UTC)

I have begun investigating support for the traditional script for the Lü language of Yunnan, China. ISO language code 'khb'

The modern script, New Tai Lue (ISO code 'TALU') is working fine if you have a font installed. The CSS in Medaiwiki:Common.css is set to use any of several known fonts and it works with both {{l}} in the various links and {{t}} in translation tables.

I've just added some known working fonts for the traditional script, Lanna aka Tai Tham (ISO code 'LANA'). This seems to work fine for {{l}} links but is not working for {{t}} translation entries.

See both scripts working for {{l}} links: Wiktionary:Requested entries (Lü)

See the {{a}} translation entry working for 'TALU' but broken for 'LANA': and#Translations

For some reason, translations with language set to 'khb' and script set to 'LANA' in the resultant HTML end up with a CSS class 'None' rather than the expected 'Lana'.

I don't know the full pipeline of templates, modules, CSS files etc that might play a part. Can anyone help get this working? — hippietrail (talk) 10:12, 23 February 2016 (UTC)

Did improve things? - -sche (discuss) 02:25, 24 February 2016 (UTC)

Yes it did! Thank you. — hippietrail (talk) 06:31, 25 February 2016 (UTC)

I removed it but it got reverted: https://en.wiktionary.orghttps://en.wiktionary.org/w/index.php?title=idem&type=revision&diff=37419993&oldid=37419944
<i.n> is a malformed html tag and can't be considered valid in the wiki syntax. 73.71.174.75 21:53, 23 February 2016 (UTC)

It is part of the syntax used by the {{la-decl-multi}} template. —JohnC5 22:02, 23 February 2016 (UTC)

This is a poorly designed template because it uses the html tags for something they aren't designed for. There is only a fixed list of html tags that wiki officially supports. HTML in wikitext, and there is also the detailed official documentation somewhere specifying which tags are allowed, and which ones are considered self-closed and which ones aren't. 73.71.174.75 22:20, 23 February 2016 (UTC)

Before you go calling Kenny's template poorly designed, I'd point out that the angle brackets are not html tags at all but an input syntax used internally by the module. They are not and would never be rendered as html tags by the html parser. —JohnC5 23:38, 23 February 2016 (UTC)

They are present in the wiki document, therefore such syntax implicitly exploits the lack of syntax checking in the wiki parser, which is never a good thing (to exploit some obscure and undocumented things). If the parser checked the syntax better, this would have been a syntax error that it really is. 73.71.174.75 02:40, 24 February 2016 (UTC)

I don't see how it's "exploiting" anything. The angle brackets are part of the input to the template, and they get processed before the final page is rendered. They are not "syntax" targeting the "wiki parser". Equinox ◑ 09:15, 24 February 2016 (UTC)

<author here> Any better methods are welcome. --kc_kennylau (talk) 11:07, 24 February 2016 (UTC)

</author here> just in case parser screws up.--Giorgi Eufshi (talk) 12:30, 24 February 2016 (UTC)

LOL. @Kenny: Maybe use parentheses or square or curly brackets instead? - -sche (discuss) 17:52, 25 February 2016 (UTC)

<no, parser will not screw up> @-sche: You can help doing it. --kc_kennylau (talk) 15:07, 27 February 2016 (UTC)

I think Template:langcatboiler should take a parameter like country=-, for use with proto-languages and constructed languages (and possibly old languages like Thracian), which would stop them from categorizing into Category:Languages not sorted into a country category. - -sche (discuss) 02:24, 24 February 2016 (UTC)

In the interim, I've added constructed languages to Cat:Languages of the world. - -sche (discuss) 22:12, 27 February 2016 (UTC)

Concerning the foreign WOTD проплывать: This verb is composed of prefix про- plus combining form -плывать; there are many other related verbs with the same comining form and other prefixes, and many more verbs related to the corresponding perfective verb проплыть. All are listed under плыть. I'd like to put an entry under "====Related terms====" that says essentially

• See плыть (plytʹ) for related terms.

Are there any templates that do it? If not, what's a good name for a template to express this? Maybe {{related-see}}? Benwing2 (talk) 08:39, 24 February 2016 (UTC)

How about transcluding плыть/Russian/Related terms? --Giorgi Eufshi (talk) 08:53, 24 February 2016 (UTC)

Hmm, how do you do that? Benwing2 (talk) 00:04, 25 February 2016 (UTC)

You do know that every page can be transcluded, rite?. --Giorgi Eufshi (talk) 06:08, 25 February 2016 (UTC)

If you create a page ], you can transclude it on another page as if it were a template by writing {{:плыть/Russian/Related terms}}. (Be sure to remember the leading colon.) —Aɴɢʀ (talk) 08:36, 25 February 2016 (UTC)

I'd prefer to direct users to the entry where the content is, rather than shifting the content onto a 'subpage' which is from a technical standpoint still a page in the main namespace (just like s/he is not a subpage of s) that will show up in search. I see "see foo" (untemplatized) a lot. We could create a template. (Maybe even repurpose {{see}}? Or use {{vide}}?) - -sche (discuss) 09:03, 25 February 2016 (UTC)

I'm surprised that there's apparently no language code for Old Czech. There's Old Polish, zlw-opl, which suggests that Old Czech should be zlw-ocs. Can someone create this? Benwing2 (talk) 00:07, 25 February 2016 (UTC)

@Benwing, Benwing2: done. —JohnC5 03:51, 25 February 2016 (UTC)

Thanks! Benwing2 (talk) 03:52, 25 February 2016 (UTC)

Thanks, this has always bothered me. Although I had assumed there was a good reason for its absence, but I have learned not to assume such things anymore. --WikiTiki89 16:08, 26 February 2016 (UTC)

Why does Category:Old Armenian words prefixed with ան- (on) give an error message, whereas Category:Old Armenian words suffixed with -ան (instrument-suffix) doesn't? --Vahag (talk) 09:15, 25 February 2016 (UTC)

@Vahagn Petrosyan: Because of https://en.wiktionary.orghttps://en.wiktionary.org/w/index.php?title=Template%3Aprefixcat&type=revision&diff=37427271&oldid=37101247 . --kc_kennylau (talk) 13:13, 25 February 2016 (UTC)

I see, thanks. --Vahag (talk) 14:59, 25 February 2016 (UTC)

And why does the category title say "on", and the "id=" parameter say "on", but the automated transliteration say "...the prefix ան- (an-)."? - -sche (discuss) 17:49, 25 February 2016 (UTC)

What appears in the category name is the id, it's used to disambiguate the category when there are multiple homonymous affixes. So it's not a transliteration. —CodeCat 18:06, 25 February 2016 (UTC)

Oh! Useful. - -sche (discuss) 18:45, 25 February 2016 (UTC)

Another useful feature would be if the category description "Old Armenian words beginning with the prefix {{m|xcl|ան-}}" generated by {{prefixcat}} would automatically be "Old Armenian words beginning with the prefix {{m|xcl|ան-|id=on}}". --Vahag (talk) 19:05, 25 February 2016 (UTC)

Done. I thought I already did it when I added the id feature, but I must have missed it. —CodeCat 19:20, 25 February 2016 (UTC)

It's working, thanks. --Vahag (talk) 19:42, 25 February 2016 (UTC)

Hi. I was just congratulated on performing my 100th edit. That's pretty lame, IMHO. Do we get new congratulations on 1000th edits, millionth, etc. ? I suppose I could avoid it but changing usernames every 99 edits...--Sit comfy (talk) 08:42, 26 February 2016 (UTC)

Where did you receive these congratulations? Your user talk page has never been edited. —Aɴɢʀ (talk) 08:54, 26 February 2016 (UTC)

Was it a "Thanks" perhaps? I see you received one of those, they are manually done by any user who feels like saying thanks to you for something you did. If you don't want to see them you can turn off notifications for "thanks" in your preferences. I am not sure that you can prevent them from being sent to you entirely. - TheDaveRoss 12:05, 26 February 2016 (UTC)

I think it's an automatic notification by the system. I got something similar after 10 edits while fooling around on the Test Wikipedia. —suzukaze (t・c) 12:15, 26 February 2016 (UTC)

I think that MW wants to make wiki editing less hostile for newbies. Because they can't change human behavior, they are trying to use the software in support of the objective. DCDuring TALK 18:16, 26 February 2016 (UTC)

Must be some automatic over-user-friendly software thing. It's not so bad, I guess. --Sit comfy (talk) 10:20, 27 February 2016 (UTC)

An index would be nice. Donnanz (talk) 13:15, 26 February 2016 (UTC)

I'm not quite sure what you're looking for, but I added {{en-categoryTOC}} in case that's what you wanted. —Aɴɢʀ (talk) 15:31, 26 February 2016 (UTC)

Yep, that's it. Cheers! Donnanz (talk) 15:47, 26 February 2016 (UTC)

This template was made to autodetect the type of morpheme based on where hyphens are placed. A hyphen at the start makes it a suffix, at the end makes it a prefix, on both ends makes it an interfix, and no hyphens means a regular word. This works well enough in practice that several editors now use it by default. But it has some limitations: not all words beginning or ending with hyphens are prefixes and suffixes. Proto-Indo-European roots, for example, are not prefixes but should be treated as regular words. Some PIE suffixes like *-éye- "look" like interfixes because they have a hyphen at both ends, but the hyphen on the right is really a stem termination. The Finnic languages also have a few cases where prefixes are used as roots for word formation. For example, Veps edeta uses ezi- as the root, and adds a suffix to it. Using {{affix}} would not work here, as it would consider ezi- a prefix when it's not in this case.

I would like to know what others think about this. How could we improve the template so that it allows it to be used in these cases as well? I can think of two possible solutions myself:

• Use another mark instead of - to indicate a hyphen that is not indicative of morpheme status. We could use ~ or -- for example: {{affix|vep|ezi--|-eta|id2=neb}} in the Veps example above. The template would automatically convert it to a proper hyphen before displaying and linking.
• Use a parameter such as affixtypeN= to override the automatic detection. The Veps example then might become {{affix|vep|ezi-|affixtype1=-|-eta|id2=neb}}, with - indicating that it's not an affix.

—CodeCat 18:09, 26 February 2016 (UTC)

I had encountered this issue before. I think I like the first option, if I've understood it correctly. So *tep- + *-éye- would be {{affix|ine-pro|*tep--|*-éye--}}? —JohnC5 20:52, 26 February 2016 (UTC)

Yes. My proposals aren't exhaustive though, I don't think my suggestions are all that good and I'm asking for better ideas if anyone has any. —CodeCat 20:54, 26 February 2016 (UTC)

Yeah, I'm also not wild about either solution, but nothing better is coming to me at the moment. —JohnC5 20:59, 26 February 2016 (UTC)

We could go back to {{compound}}, {{prefix}}, and {{suffix}} for such cases. For example, {{suffix|ine-pro|*tep-|*éye-}}. I think that's the cleanest solution. --WikiTiki89 21:22, 26 February 2016 (UTC)

That defeats the purpose of having a unified template. —CodeCat 21:33, 26 February 2016 (UTC)

No it doesn't. The unified template is still used in the vast majority of cases. --WikiTiki89 16:38, 29 February 2016 (UTC)

All the hyphens, minus signs, and dashes look too similar, there are too many of them, and only one is easily available on most keyboards. It's just too easy to use the wrong one and wonder why certain entries aren't working right with the templates. Chuck Entz (talk) 21:48, 26 February 2016 (UTC)

The proposal only involves the hyphen-minus and the tilde, both of which are ASCII characters and easily available on most keyboards. —Aɴɢʀ (talk) 21:56, 26 February 2016 (UTC)

You are, of course, right I was going back to strike my comment once I re-read the thread, but you beat me to it. Chuck Entz (talk) 22:01, 26 February 2016 (UTC)

If you have an alternative, please do propose it! —CodeCat 22:05, 26 February 2016 (UTC)

You could put a backslash before the hyphen to quote it, Unix-style -- or put the whole morpheme in quotes, which might be the best solution. Benwing2 (talk) 00:46, 27 February 2016 (UTC)

I like the idea of using a backslash. It's not a newly invented syntax but re-use of what's already commonly used and therefore more likely to be understood intuitively. It's also very unlikely to appear legitimately in any affixes. —CodeCat 01:46, 27 February 2016 (UTC)

I have been trying to link dag and herlig to the Bokmål phrase for en herlig dag! , but it doesn't work. Is the exclamation mark the problem here? I have revised the phrase entry, which links OK to herlig and dag, but not the other way round. Donnanz (talk) 11:57, 27 February 2016 (UTC)

{{l}} strips question marks and exclamation marks in terms of links (but not display) per WT:Entry titles. I think this answers your question. Renard Migrant (talk) 12:23, 27 February 2016 (UTC)

So you redirected the page (minus !) and restored the header. I hope I don't come across too many of those. Thanks a lot! Donnanz (talk) 12:45, 27 February 2016 (UTC)

There seems to have been a recent change to something with the result that when Gothic words are enclosed by {{l}}, the automatically generated transliteration is now automatically also linked. In other words {{l|got|𐍈𐌰𐍃}} now results in 𐍈𐌰𐍃 (ƕas) with the "ƕas" as a blue link, instead of as unlinked black text, as it used to be. This is a good thing, since we have entries for all attested Gothic words in their romanized forms, so why not link to them. However, the link is also there even when the Gothic word's own link is overridden by being put in position 3: thus {{l|got||𐍈𐌰𐍃}} results in 𐍈𐌰𐍃 (ƕas) with the "ƕas" as a blue link, even though "𐍈𐌰𐍃" itself does not have a link. Could this be changed, so that the romanization is linked only if the native script version is also linked? The reason is that I use {{l}} when providing quotations, putting the Gothic text in position 3 since we don't want a link to a whole sentence, but then the transliteration of the entire quotation is rendered as a (red) link: see for example the quotations under the definition at 𐌰𐍆𐌼𐌰𐍂𐌶𐌾𐌰𐌽. —Aɴɢʀ (talk) 18:21, 27 February 2016 (UTC)

I think this should be fixed, in general. But the problem at 𐌰𐍆𐌼𐌰𐍂𐌶𐌾𐌰𐌽 (afmarzjan) was fixed by using {{ux}}, which is meant to be used in these cases. —CodeCat 18:28, 27 February 2016 (UTC)

{{ux}} is for usage examples, not quotations, isn't it? And anyway, 𐌰𐍆𐌼𐌰𐍂𐌶𐌾𐌰𐌽 isn't the only one: I've added quotations to a whole slew of Gothic words using {{l|got||...}}. —Aɴɢʀ (talk) 18:36, 27 February 2016 (UTC)

Quotations are a kind of usage example. —CodeCat 18:46, 27 February 2016 (UTC)

At least in Latin-alphabet languages, they're formatted differently. Usage examples are italicized, quotations aren't. —Aɴɢʀ (talk) 18:55, 27 February 2016 (UTC)

Well then, use a template equivalent to {{ux}} but intended for quotations. —CodeCat 20:27, 27 February 2016 (UTC)

I agree. I've used a bot to rewrite cases where {{l}} is being used for quotations to use {{ux}} (or {{ru-ux}}, which probably should have its special features merged into {{ux}}). I think the docs saying that {{ux}} shouldn't be used for quotations should be changed. Benwing2 (talk) 21:25, 27 February 2016 (UTC)

We have two templates for quotations: {{Q}} and {{quote-book}}. {{ux}} cannot do their job. --Vahag (talk) 21:29, 27 February 2016 (UTC)

Even if I stop using {{l}} for quotations (and I've been using it for quotations in all languages other than English for quite some time now), I still feel that if a term is put in pos.3 of {{l}} in order to override the link, then the link to the romanization ought to be overridden as well. —Aɴɢʀ (talk) 22:22, 27 February 2016 (UTC)

I agree. For example, someone might want to mention (and correctly tag, etc) a hypothesized Gothic word in an etymology, for example (citing some scholar who suggests the term under discussion is from that Gothic word). - -sche (discuss) 22:27, 27 February 2016 (UTC)

Does anyone know a way to find out which {{R:TLFi}} templates link to an error 404 page like this one (fractal). Kennybot added {{R:TLFi}} to all pages in Category:French lemmas and rather obviously some of those pages don't have corresponding TLFi entries. Renard Migrant (talk) 21:54, 27 February 2016 (UTC)

Why is this happening, esp. as there is no consensus to do this? These two are synonyms, what is the point of renaming to a longer name that adds nothing? Benwing2 (talk) 21:59, 27 February 2016 (UTC)

Presumably it's preparation for using {{quote}} for non-block-quote quotations, as a parallel to {{ux}}. It is unintuitive that "quote" currently formats things very differently from all the templates that begin with "quote" ("quote-book", etc). However, there should have been some notice and discussion about that. (Tangential: I find it unhelpful that both {{q}} and {{Q}} exist and do very different things; our templates are usually only case-sensitive in terms of whether or not they exist, and not in terms of what they do.) - -sche (discuss) 22:11, 27 February 2016 (UTC)

Both things are addressed at WT:RFM, where I created a move proposal. —CodeCat 22:15, 27 February 2016 (UTC)

I was preparing quote for inline quotations but I think Q->quote rename has more merits. --Dixtosa (talk) 22:21, 27 February 2016 (UTC)

I did see your WT:RFM proposal but (a) it doesn't mention renaming quote -> blockquote, and (b) IMO you should wait for the proposal to be discussed before going ahead with a rename like this. Bot changes are supposed to be uncontroversial. Benwing2 (talk) 22:29, 27 February 2016 (UTC)

It was uncontroversial before anyone complained. In any case, I didn't move {{quote}}, Dixtosa did. I just edited out the redirects. —CodeCat 22:35, 27 February 2016 (UTC)

OK, didn't realize that Dixtosa moved it. Benwing2 (talk) 01:19, 28 February 2016 (UTC)

Example: Category:Serbo-Croatian twice-borrowed terms. Instead of being one or two lines, the table of contents is spread out over several lines:

A B C Č

Ć D Dž Đ

E F G H

I J K L

Lj M N

Nj O P R

S Š T U

V Z Ž

А Б В Г

Д Ђ Е Ж

З И Ј К

Л Љ М Н

Њ О П

Р С Т Ћ

У Ф Х Ц

Ч Џ Ш

Is this intentional or a bug? It's ugly and pushes the category contents off the screen, which is bad for usability. - -sche (discuss) 03:13, 28 February 2016 (UTC)

See {{sh-categoryTOC}}. It looks like it's intentional. —suzukaze (t・c) 03:15, 28 February 2016 (UTC)