User talk:Trappist the monk/Archive 15

Cite Error Message

Hey- I have recently made some edits to Help:Cite errors/Cite error references no text#Issues and resolution. Let me know if it's a worthwhile addition. It may need to be re-worded. Geographyinitiative (talk) 13:42, 11 October 2020 (UTC)

I believe that what you wrote is correct. Beyond that, I have no opinion because those sorts of errors do not fall within the cs1|2 remit.

—Trappist the monk (talk) 13:50, 11 October 2020 (UTC)

Transclusion update

Hey, could you use your bot (which you used to update transclusion links after you moved the module) to update the transclusion links of Module:Language/data/wp languages? It's been almost 10 days and it still has over 200k transclusions. I'd like to be able to see where it's still in use. --Gonnym (talk) 13:24, 12 October 2020 (UTC)

Why impatient? Is there a deadline? The bot is running. It will take a while...

—Trappist the monk (talk) 13:34, 12 October 2020 (UTC)

I wouldn't call waiting 10 days impatient. If you don't want to that's fine, but no need to belittle. There might not be a deadline, but I also don't want to have to wait a year to finish a simple project when every step that could have taken 1 day, takes 2 weeks+. --Gonnym (talk) 13:44, 12 October 2020 (UTC)

What was the transclusion count 10 days ago? Updating transclusions, like populating categories, is not an instantaneous thing. Each page has to be refreshed. If there is nothing else for the servers to do, I would imagine that they could get through 200k pages pretty quickly. But, every human and bot edit, like this post, has priority over updating pages that are on the job queue because of template and module updates (looking at Special:RecentChanges, right now, humans and bots are making about 100–200 pages per minute depending on how you filter the list). The Module:Lang update dumped 1.1 million pages on the job queue; over the weekend I updated the Module:Citation/CS1 suite so 4.6 million more got added to the job queue. I don't know how pages on the job queue are prioritized. Lots to do and it all takes time...

—Trappist the monk (talk) 14:17, 12 October 2020 (UTC)

Your revision of the Cite Jewish Encyclopedia template

One of the changes I noticed that you made in your revision of Template:Cite Jewish Encyclopedia is that previously the wstitle parameter would have precedence of the title or article parameter, while now it is the other way around. Why did you make that change? Debresser (talk) 21:11, 11 October 2020 (UTC)

I'm having trouble decoding what it is that you are asking. Can you restate the question or show an example where the template does something that it should not do?

Module:Citation/CS1 now detects empty parameters that are not known to it so that such detritus can be removed. Before my edit, this template, in common with several of the other encyclopedia-wrapping templates, created |HIDE_PARAMETERn= parameters that get passed through {{cite encyclopedia}} to Module:Citation/CS1. These 'hidden parameters' – they were never really 'hidden' – are not known to the module so it emits error messages and puts the articles that use these templates into Category:CS1 errors: empty unknown parameters. My purpose in editing this template is to get the articles where it is transcluded out of that category. My secondary reason was to adapt this template to use Module:Template wrapper so that it is no-longer necessary to add special support for new or revised cs1|2 template parameters – any parameter normally acceptable to the wrapped template, unless excluded, passes directly to the wrapped template. I have done this with several of the encyclopedia-wrapping templates.

—Trappist the monk (talk) 22:09, 11 October 2020 (UTC)

I see that you actually did understand what I meant.[1]

You did a great job, and I love how things are a lot simpler now. Debresser (talk) 22:23, 13 October 2020 (UTC) Debresser (talk) 22:23, 13 October 2020 (UTC)

Help with regex and the Meta discussion

Hey, Trappist!

I wanted to ask you a silly regex question regarding AWB: I have around 200 category pages the content of which starts with 2 newlines. I want to remove them. How do I do that with AWB? I thought a F&R "\n\n" -> "" would deal with it but apparently it doesn't, strangely enough.

I also wanted to give you the results that came from that discussion on Meta regarding the globalization of CS1 module. Unfortunately, it didn't happen. I tried two times to elicit some kind of discussion but I got no answer. :/ - Klein Muçi (talk) 16:58, 13 October 2020 (UTC)

Link to one of these cats?

—Trappist the monk (talk) 17:01, 13 October 2020 (UTC)

All the pages of the subcategories listed here.

Also, if you're interested, my 2nd try on Meta: Here. - Klein Muçi (talk) 17:05, 13 October 2020 (UTC)

Because it appears that the text in every category begins with the same word, try:

f: [\r\n][\r\n]K

r: K

But, why? If you are going to go to the trouble of removing the two newlines, why not replace the content with a template? Those same categories here use {{CS1 language sources}} and {{CS1 language sources/core}}so that the documentation is held in one place and customized for each category according to the category's name.

The regex to do that would be, I think, multi step: find and replace all [\r\n]+ with empty strings to convert the category text into a single string. Find everything except the category link and replace all of that with empty string (f: ^[^\[]+). Should leave only the category link, so replace that with the template call.

Can't say that I'm surprised at the boundless indifference you experienced at meta...

—Trappist the monk (talk) 17:47, 13 October 2020 (UTC)

Have you had bad experiences with Meta at the past?

As for the regex subject... The thing is SqQuote is far behind SqWiki and EnWiki in template terms. The very reason those 2 newlines are there it's because of the removal of a template. I imported all the categories from SqWiki and there we use a template indicating that those categories are maintenance categories. It's an old, outdated template which in SqQuote doesn't exist. Not wanting to just make a problematic copycat in SqQuote, I removed that template with JWB which basically replaced it with a blank line. Add the newline the categories had itself to split the template from the text and you get the 2 newlines I'm trying to remove now. I would go on with your proposal but unfortunately I have most important things I need to attend to like importing around 40 more categories related to CS1, investigating what has caused Smallem to stop working and updating the whole module again in both projects because I noticed it was falling behind EnWiki again in terms of functionality, a thing which, unfortunately I've come to dread because of the many bugs that arise after that process. Unfortunately even what you suggested doesn't work. Everything gets skipped and nothing gets changed. I tried playing around with the options "multiline", "singleline", "regex" and their combinations but nothing changes. I even tried looking at the skip options if anything wasn't like it should be but I couldn't find any problems. Any idea what I might be doing wrong? :/ - Klein Muçi (talk) 18:13, 13 October 2020 (UTC)

Never had any dealings with meta that I can recall. You are proposing a big project that isn't at all glitzy and will require a substantial amount of work... It is the unique individual who will sign up for that.

When I copy the content of sq:Faqe me burime në abkazisht (ab)‎ into the Text to search box in the awb regex tester, set Find to: [\r\n][\r\n]K, Replace to K, uncheck all of the RegexOptions, and then click Find, it shows me something that looks sort of like this:

• {\n\nK}

  {\n\nK}

When I click Replace, the Results box shows the category's text without the two leading newlines. This same should work for you.

—Trappist the monk (talk) 18:36, 13 October 2020 (UTC)

Oh, I see now...

As for AWB, unfortunately I couldn't and still can't make it work in AWB. The program just skips every page from beginning to end until it empties the list without doing any changes. I tried it with a random Find and Replace and still did the same thing so it's a problem unrelated to this subject. But I was able to make it work in JWB with your regex so it's cool. :) - Klein Muçi (talk) 02:36, 14 October 2020 (UTC)

Staying on the same subject, is there an easy way I can find what's here but not here? - Klein Muçi (talk) 03:15, 14 October 2020 (UTC)

Nevermind, I found them. The problem is that I'm unsure what to do with them now. One is the Albanian one and of course, it doesn't appear in the foreign languages for us. The other 4 are these:

• Category:CS1 Taiwanese Mandarin-language sources (zh);
• Category:CS1 Oriya-language sources (or);
• Category:CS1 Haitian Creole-language sources (ht);
• Category:CS1 Valencian-language sources (ca);

They all have ISO codes that are already used by some other languages, in that sense they appear as duplicates (the only ones to do so in that list) and I just remembered that's why I hadn't created those as categories in the first place. Can you explain to me a bit more in detail what's happening and if those categories should also be created if I intend to "cover everything"? (I mean, given that I've already created 185 categories, 4 more are nothing.) - Klein Muçi (talk) 11:28, 14 October 2020 (UTC)

Taiwanese Mandarin and Oriya are not the MediaWiki names associated with zh and or; you will not find them in the list of supported languages which see. Haitian Creole is the MediaWiki supported language name for ht; Haitian is not:

{{#language:ht|en}} → Haitian Creole

I will delete those three categories. Thanks for pointing them out.

Catalan and Valencian share the same root ISO 639-1 code ca. Valencian is not a MediaWiki recognized language but there was some agitation here when |language=Valencian categorized as an unrecognized language so cs1|2 supports the IETF tag ca-valencia.

When you tried to use awb to replace the two newlines did you have the Only whitespace checkbox checked on the Skip tab?

—Trappist the monk (talk) 12:03, 14 October 2020 (UTC)

Oh, so I should use Haitian Creole instead of Haitian? As for the other 2 languages above that, I guess I'm using the "right version" since you said you were deleting them. As for Valencian/Catalan... From my perspective, the perspective of someone that has the only guess (not knowledge) that Catalan may be a language spoken in the Iberian Peninsula and knows nothing about Valencian... I guess I should use only one? :/

No, I didn't. I unchecked everything from the Skip tab. I even tried it once with each one of those 3 last options in the end. Exists, Doesn't exist, Don't care. Nothing changed. :/ - Klein Muçi (talk) 13:18, 14 October 2020 (UTC)

Did you look in the Logs tab to see if awb gave a reason for skipping? The categories are protected so perhaps they were skipped because of that?

Yes, Haitian Creole. Unless you remove the cs1|2 Valencian override, I would include the Valencian category because this:

{{cite book |title=Title |language=ca-valencia}}

will categorize at:

sq:Category:Faqe me burime në Valencian (ca)

—Trappist the monk (talk) 13:44, 14 October 2020 (UTC)

Okay, done. As for AWB, I redid it now and the logs give this reason: Skipped by custom module () What does it mean? :/ - Klein Muçi (talk) 14:21, 14 October 2020 (UTC)

Umm, sq:Category:Faqe me burime në valencisht (ca)‎ is not the name of the category that cs1|2 populates. To make cs1|2 populate that category you need to change sq:Module:Citation/CS1/Configuration.

In awb: Tools > Make module; Enabled is checked, right? Whatever you have there is setting the Skip return boolean to true likely because whatever the module is looking for, it did not find.

It's a good idea to start a fresh instance of awb whenever doing these little jobs like the \n\n removal. You might create and save a blank default settings file to use as a starting point for these kinds of jobs.

—Trappist the monk (talk) 14:37, 14 October 2020 (UTC)

Yes, it was checked ever since that last time you gave me that module to use. It works now! Thank you! <3

Regarding the change in CS1 module, is this the line I need to change: ['ca-valencia'] = 'Valencian' ? From Valencian to Valencisht? - Klein Muçi (talk) 15:19, 14 October 2020 (UTC)

Two places. The other is:

['valencian'] = {'Valencian', 'ca'}

to:

['valencisht'] = {'Valencisht', 'ca'}

—Trappist the monk (talk) 15:28, 14 October 2020 (UTC)

Done, I hope. - Klein Muçi (talk) 15:47, 14 October 2020 (UTC)

Nope. ['valencian'] → ['valencisht']

—Trappist the monk (talk) 15:51, 14 October 2020 (UTC)

Take a look at our module. I guess now it's all right, no? - Klein Muçi (talk) 16:22, 14 October 2020 (UTC)

Yeah, but, shouldn't all of the names on the left in lang_name_remap{} be Albanian-language names? Was there a reason that you left them as English names?

—Trappist the monk (talk) 16:31, 14 October 2020 (UTC)

Yes: Ignorance. :P I didn't know how to convert them. We've had a hard time converting many of the 185 other languages in Albanian. We don't have names for many of them and we've had to create them on the go. Usually just by getting the stem of the noun and adding -isht at the end of it, which is how many of the language names in Albanian are formed. This has created many hard-to-pronounce names and some rather funny ones but we don't know what to do with that (I've talked with some people). Some times, a further study of the said languages could bring better translated names (if we know what we are dealing with) but no one so far has done that. Given that those ones you mention are very rare occurrences for us, I've left them in original until I'm forced to deal with them, like I just did with Valencian. One rather problematic translation was Blackfoot, because the names have meanings and I don't know if we should preserve those meanings or how to do that, if we should. Other exemplars there too have their difficulties in translation. - Klein Muçi (talk) 17:04, 14 October 2020 (UTC)

For the more obscure languages I guess I'm not surprised that there isn't an Albanian name. Probably best to keep the English names in those cases because at least those names are tracible? Blackfoot would be pretty obscure by that name or by its native (and ISO 639-preferred) name: Siksika.

—Trappist the monk (talk) 17:33, 14 October 2020 (UTC)

Maybe Siksika would be a bit better in Albanian translation because we could say siksikisht. That's what I meant with a study. Some of them could be converted a bit better if we could learn a bit more about them. But anyway, for now we're good. :) - Klein Muçi (talk) 18:09, 14 October 2020 (UTC)

Technical advice requested

I see by this edit you did that apparently it's best to leave out {open access} and ref=harv. I think you have more technical knowledge about this than I do. Thanks for advice.--Doug Coldwell (talk) 22:09, 16 October 2020 (UTC)

The term open access really means that the resource is available to be used in any way you want as long as you have an agreement with the publisher to do that. Here at en.wiki, what editors generally mean when they use {{open access}} is that the resource is free-to-read. The cs1|2 citation templates follow the conventions that anything linked through a title-holding parameter is free-to-read unless marked otherwise; conversely, named identifiers are presumed to lie behind a paywall or registration barrier. Details at Help:Citation Style 1 § Registration or subscription required.

|ref=harv is no longer needed; the function that it once performed is now the default.

Did I answer your questions?

—Trappist the monk (talk) 22:35, 16 October 2020 (UTC)

Perfectly. Now I understand, as I had been using those for years before. Now I get it. Thanks for info.--Doug Coldwell (talk) 09:06, 17 October 2020 (UTC)

Public discussion about templates

Hi!

Everybody names you as the #1 expert on citations templates and modules, so I'm asking you :)

Next week there is an online meeting about Wikicite. I am looking for ways to bridge the futuristic ideas of Wikicite with the actual realities and technicalities of editing the wikis. I proposed a session about this, and I'd love to hear more expert voices there. Here is the what I wrote about my presentation meta:WikiCite/2020 Virtual conference/The frontend of WikiCite.

Would you be willing to join the panel discussion, and perhaps present some of your own ideas, or challenges of which you can think in transition to any future systems?

Thanks! --Amir E. Aharoni (talk) 18:45, 19 October 2020 (UTC)

Thank you, but I am going to decline your invitation because panel discussions are not my métier. If anyone thinks it useful, I am happy to answer questions.

—Trappist the monk (talk) 19:34, 19 October 2020 (UTC)

Thanks for the quick response! :)

Will you be comfortable with being in the call, and answering questions if they come up? Or are you only comfortable with written questions? --Amir E. Aharoni (talk) 19:37, 19 October 2020 (UTC)

Nowhere near as fast as yours... Written. I like to think about what I should say before I say it; if I should say it.

—Trappist the monk (talk) 19:46, 19 October 2020 (UTC)

No problem, thanks :) --Amir E. Aharoni (talk) 20:41, 19 October 2020 (UTC)

Limit bot to main space

Any page outside of main space may contain intentional errors as examples. Therefore edits such as one in help: space is inappropriate. Jc3s5h (talk) 07:42, 20 October 2020 (UTC)

Smallem

Short, direct question: Can Smallem help solve problems in Category:CS1 errors: empty unknown parameters by simple Find and Replace regular expressions? I've read the category's description 3 times now and still I'm not sure what cases it covers. In my head it feels like its functionality is already covered by other existing error categories. And I'm not really sure about the practical cases, as I said. For example, would it catch cases when every parameter is between two pipes, which makes the end of the citation appear like this: ...|last-parameter=last-value|}}? That's a practice that is not unusual to do. Is the whole category concerned (practically) only with the pipes? Because if it is, then maybe Smallem can help with it but I doubt that's the case. - Klein Muçi (talk) 11:34, 21 October 2020 (UTC)

The purpose of the category is to see just how much junk is out there.

The empty positional parameters can be found and removed with something like this:

f: \|(\s*[\|\}])

r: $1

Do not use that regex as-is; you must apply it only to cs1|2 templates because empty positional parameters are allowed and used in all sorts of other templates.

Empty named parameters where the name is not known to cs1|2 is a harder nut to crack because smallem will need to keep a list of the names of all cs1|2 parameters. It can then look look in the list to see if |blue= is a valid cs1|2 parameter; it's not so can/should be deleted. Alternately, what I would do, were it possible at en.wiki without running afoul of WP:COSMETICBOT, is delete all empty parameters in every cs1|2 template. When they are empty they serve no purpose; the notion that if-I-leave-this-empty-I-or-some-other-editor-will-fill-it is a myth.

—Trappist the monk (talk) 12:10, 21 October 2020 (UTC)

So basically what I mentioned above as a practical example is caught up in that category and my overall understanding of it is correct. As a side-self-note, apparently I have problems with my module copy at SqWiki, because I'm well aware many cases where what I gave as an example is used in our project and right now there are no articles at that category. That's because, that used to be my method of adding citations in my early wiki days. I kept every parameter between 2 pipes to help my future-self in case I discovered more parameters that could be added to templates and their accompanying information (I wasn't aware of the full lists that exist in cite template's documentation's pages back then). Staying on the same time frame, I also thought that the correct way to fill in references was to put every parameter possible in a template, like we usually do with infoboxes, and leave what we can't fill empty. The only reasons I didn't do that is because at first I wasn't aware of the said lists of parameters and when I was made aware of them, I saw they were gigantic and this came at a time I had already started to dislike citations in general because of the cluttering they made to the text and the alienation that came with it from a normal text to a technical script-looking one. So I know what you mean and apparently that's a pretty common way of thinking. Okay then, thank you for your answer!

Given that I was inspired by my old memories, take what I'm saying as food for thought with more than a grain of salt, but do you see it as possible, maybe in a distant future, :P to have citations in a kind of different "namespace"? Like we have plain article text at one side and citations on another and we just put incrementing numbers where we want them on the "plain" article text: [1] = first citation in the row, [2] = second citation in the row and so on. One might say why only citations and not all templates (and I guess that's what VE strives for in a sort of way) but the problem with citations is that they go in the middle of the sentences making the text very hard to read in edit mode if you haven't syntax highlighter activated. Other templates rarely, if ever, go in the middle of sentences. Maybe what I'm proposing falls in the program of WikiCite and their database? - Klein Muçi (talk) 12:47, 21 October 2020 (UTC)

A pipe introduces a parameter; it does not close a parameter so in |}, the pipe introduces an empty positional parameter.

List defined references is one way to avoid the clutter. Editors will always need some way to say put-this-citation-here whether is is LDR and self-closed named <ref name="..." /> tags, {{sfn}}, or some other template or tag. I don't think that the WikiCite initiative will change that except that the data for the citation will be stored as wikidata and not be in that article wikitext at all.

—Trappist the monk (talk) 13:26, 21 October 2020 (UTC)

LDR looks more like what I was mentioning but it still doesn't look minimalistic enough in its approach (at least, compared to what I proposed above) and it's not VE complaint apparently so that would make it to be a bit frowned upon new users, which I'm trying to make happy in the first place. But it's good that there have been initiatives on that direction and I'd be glad to see them popularized in the future, even though for me maybe it would make it harder to work with considering I've been forced to learn "the old school" way. I'm a bit confused by your last sentence. If they become stored in Wikidata, and this approach becomes the default one, (maybe this is the part you're not taking for granted, not that you should) wouldn't the end result be practically what I explained above, articles with no citation clutters?

It's a good thing you gave me that definition about pipes because it will help in explaining those in workshops/edit-a-thons. Apparently my module is fine (just checked) but the server hasn't yet loaded all the needed pages in the category page. Staying on the same subject, since we're addressing concepts like positional parameters and empty ones, things that tend to fall in the cosmetic side, wouldn't it be wise to address and standardize even the spaces between cite-template's elements? Brackets, parameters, equal signs, values and pipes? Like, I understand people that decide to add 1 space between each element, maybe to make it easier for the eye. But adding many spaces in a row (I've seen some extreme cases with more than 15 spaces) or writing citations vertically (pressing Enter instead of Space) are strange to see and make the whole text layout difficult to navigate with. - Klein Muçi (talk) 13:49, 21 October 2020 (UTC)

There will always need to be some sort of marker in wikitext to show where a citation should be placed. You used [1] and [2] as examples; LDR uses <ref name="..." /> tags. WikiCite may use something else though I expect that it will be something like <ref Q="<digits>" /> or some such. This is an implementation detail that need not be solved right now. And rightly so. Because there will be bazillions of citations in wikidata, the hard task is to make it easy for editors to quickly and accurately locate the one citation from all of those bazillions to insert into an article all the while ensuring that the citation points to the source that the editor consulted. How it gets inserted and what that insertion looks like is comparatively trivial.

Yeah, uniform parameter style is another one of those were-it-up-to-me things. Space characters help with line wrapping and readability. If everyone believed that, there would not be the large number of unspaced cs1|2 templates...

—Trappist the monk (talk) 14:42, 21 October 2020 (UTC)

I understand. I just hope they have a minimalistic approach so as it doesn't disrupt normal reading. As for the parameter style, what do you mean practically? Do you think it should or shouldn't be standardized/uniformized, were it up to you? Personally I usually use 1 space character between elements in wikimarkup for readability except for citations. Unfortunately there, my frustration from "text clutter" forced me to try and remove as much as I can as time went on. But I can agree one 1 space. Do you agree on that or do you think that it shouldn't be standardized? The reason I'm asking it's because if we agree on that point, maybe you could "experiment" on SqWiki with that concept. Creating an error related to that aspect and an accompanying category for it. Here called an "experiment" because it would be a first, I believe it would be more than welcomed in our community. Things change though if you have a difference stance related to it, in which case you could explain to me why something like that might be wrong.

I've asked you something on the end of the discussion we had prior to this. Maybe it has escaped your notice. - Klein Muçi (talk) 15:17, 21 October 2020 (UTC)

My preferred style is |param-name=value |param-name=value. But, we all have preferred styles. What I am opposed to is |param-name=value|param-name=value especially when there aren't any hyphen or space characters in parameter name and paramter values:

|last1=name|first1=I|last2=name|first2=I|last3=name|first3=I|last4=name|first4=I|last5=name|first5=I etc ad nauseum

That style offers no possibility for rational line wrapping.

If you are suggesting that cs1|2 should emit an error when various spacing (or non-spacing) is used in a raw template, not going to happen. MediaWiki reads the template and hands an arguments table of key/value pairs to Module:Citation/CS1 that looks like:

['key'] = 'value'

where key is the parameter name and value is the parameter's assigned value. In the translation from raw cs1|2 template to lua arguments table, parameter names and assigned values are stripped of leading and trailing whitespace so the module never sees how the cs1|2 template was written in wikitext.

Presumably you could train smallem to standardize cs1|2 template style if that sort of thing is permitted on your wikis.

—Trappist the monk (talk) 15:52, 21 October 2020 (UTC)

Basically you don't like my preferred style. :P But I'm starting to gravitate towards your style after reading that pipes introduce parameters, don't separate them (at least the definition, don't mind the practical function). All I wanted to "fight" though was cases like this:|param-name =value | param-name = value |param-name=value (check the wikitext layout, not its rendering)

I've seen a lot of times a crazy number of spaces between elements and in some of our old articles, from the days Wikipedia had just started, you can find all citations laid out on a vertical format, maybe inspired by infoboxes, some of the first en masse used templates on articles. But you say that the module can't see these cases so...

Considering my current knowledge on bot related things, I wouldn't know how to do that fully yet. Of course, it's easy to standardize elements related to parameters with find and replace: {{ | param-name = -> {{|param-name= - I can write many regular expressions like that to cover all templates. But when it comes to fixing the space related to values of the said parameters, that's where I get stuck, considering that values change in infinite ways and I can only work with crude regular expressions and cruder find and replace functions. - Klein Muçi (talk) 16:58, 21 October 2020 (UTC)

The dreaded time has come...

Hello! The dreaded time when I have to update the module in SqWiki came and with it came its bugs and many questions related to it.

So, first things first: I finished updating all 9 subpages. Apparently I have made a mistake somewhere (See the references here). I'll try again later to fix it even though I double checked the main page and the configuration one. Maybe your eye can catch the problem more easily?

Secondly, can you help me by telling what changes have been made to categories? What are removed (if any) and what are added new?

Thirdly, I took the courage to translate all those special languages I hadn't translated in the last conversation that we had. Can you check to make sure everything is fine there?

Also, can you tell me if I can safely remove the $1 parameter here: message = 'Cite has empty unknown parameter$1: $2'. In Albanian the "S" part doesn't make any sense and that whole code line could be rewritten differently while only using the $2 parameter. (Maybe you need to change it altogether to accommodate for I18N with different languages?)

Lastly, I would also like to ask you some questions about ID handlers but I need to solve the problems above first. - Klein Muçi (talk) 16:34, 17 October 2020 (UTC)

At sq:Moduli:Citation/CS1, change line 2598 to:

if not utilities.is_set (Language) then

Module:Citation/CS1 is getting to be too big and we ran headlong into a very hard wall so direct calls to function in Module:Citation/CS1/Utilities is necessary.

error categories:

• +Category:CS1 errors: empty unknown parameters‎
• +Category:CS1 errors: generic title‎

Maintenance categories:

• +Category:CS1 maint: DOI inactive‎
  • +Category:CS1 maint: DOI inactive as of August 2020‎
  • +Category:CS1 maint: DOI inactive as of September 2020‎
  • +Category:CS1 maint: DOI inactive as of October 2020‎
• +Category:CS1 maint: nocat‎
• -Category:CS1 maint: uses editors parameter‎

properties categories:

• +Category:CS1: abbreviated year range‎

If the language names work for you, then I have nothing to add except that someone using the English form of one of those language names won't find a match in lang_name_remap{}.

You remove $1.

—Trappist the monk (talk) 17:15, 17 October 2020 (UTC)

Can you show me what is line 2598? Unfortunately modules for me in SqWiki don't show the same as in EnWiki with numbered lines and syntax highlighting. This contributes to many difficulties in working with them. (I don't know why. :/ ) - Klein Muçi (talk) 17:21, 17 October 2020 (UTC)

if not is_set (Language) then			-- when |language= is missing or empty

—Trappist the monk (talk) 17:31, 17 October 2020 (UTC)

Also set_message becomes utilities.set_message at line 2599.

—Trappist the monk (talk) 17:39, 17 October 2020 (UTC)

Oh! Now I get it what you meant with direct calls to utilities. I fixed that second line "by intuition" after seeing the mistake. And thank you for the category list! Can you explain to me what the nocat parameter does (used to do?) And I'm a bit confused in relation to the English names of languages. Other languages (recognized normally by MediaWiki) work with the English equivalent in SqWiki, no? If so, what part do I need to leave in English to have the same results with these? Can you give me just 1 practical example? - Klein Muçi (talk) 17:52, 17 October 2020 (UTC)

|nocat= is an alias of |no-tracking=. It prevents a cs1|2 template from adding the article to a particular category. Mostly intended for use in documentation pages where it is sometimes desireable to illustrate error messages but not desireable to have that doc page listed (forever) in the error category. We would have deprecated it but had no real way of knowing how often it is used because a lot of templates that support some sort of no tracking capability use that parameter name. Category:CS1 maint: nocat‎ exists so that we can figure out how often it is used. Apparently, not much, so the category and the parameter will likely go away (no deprecation period) at the next update.

The MediaWiki language list for various non-English languages falls back (very often) to the English-language language names. For example, Cree (cr) is a North American First Nation language. MediaWiki does not have an Albanian-language name for Cree so when given the ISO 639-1 code cr and asked to return the Albanian name, it returns the English name:

{{#language:cr|sq}} → Cree

To prove that the code is working correctly, Navajo (nv) is a Native American language for which MediaWiki does have an Albanian language name:

{{#language:nv|sq}} → navahoisht

For languages that it does not override, cs1|2, looks for the value assigned to |language= in MediaWiki's list of languages for the local wiki. If I write |language=Cree in a cs1|2 template at sq.wiki, the language name will be found in MediaWiki's list. If I write |language=cr in a cs1|2 template at sq.wiki, the code will be found in MediaWiki's list and cs1|2 will render the language as Cree because there isn't an Albanian form.

cs1|2 has the override for Valencian. At sq.wiki, this:

['valencisht'] = {'valencisht', 'ca'},

That override allows editors to write |language=valencisht but fails if they write |language=Valencian because there isn't a matching entry in the lang_name_remap{} table. To support Valencian, the override table must have:

['valencian'] = {'valencisht', 'ca'},			-- translate from English

But, the real question is: Do you care? If you have absolutely nothing better to do, you might add translations from English for the overridden languages.

—Trappist the monk (talk) 18:56, 17 October 2020 (UTC)

Thank you! I will add the English terms according to your example.

Now the question related to id handlers. I have removed the 2 lines of code that make wikilinks possible in that section for every entry. The reason for that is that we don't have articles for 95% of those things yet. To not confuse users that think "red=something wrong with my reference", I removed the wikilinks (because that's what they would get with wikilinks to nonexistent articles). Is this the correct way to handle this situation? Any better suggestion? Also, why does some of those numbers there change with every update? - Klein Muçi (talk) 19:16, 17 October 2020 (UTC)

So that I don't have to guess what lines you removed, point to them?

—Trappist the monk (talk) 19:21, 17 October 2020 (UTC)

link = 'Amazon Standard Identification Number', redirect = 'ASIN (identifier)'

For example. ^ - Klein Muçi (talk) 19:27, 17 October 2020 (UTC)

I think that blunt-force removal is probably not a good idea. If you must remove those lines, comment them out with two hyphen characters at the left margin.

Which is worse? Editors think that something is wrong with the citation or readers faced with a string of indecipherable external links without a clue about what that alphanumeric soup means? For me, that's easy to answer. I would much rather have confused editors than confused readers. Citations are cryptic enough, I don't see any value in making them more cryptic. Does sq.wiki have the concept of stub articles? You could switch-on awb and relatively quickly create stub articles that, if nothing more, say that this article is a stub, suggest that interested editors might translate some-other-wiki's article. That would solve the redlink problem and possibly give your readers someplace to link if they are interested.

—Trappist the monk (talk) 20:09, 17 October 2020 (UTC)

I re-added them after almost a year being removed. We do have stub articles. How would I use AWB to make that? But I have some dilemmas. The first one being about the titles of the articles. Acronyms or full names? English or Albanian? The second one is about other projects. What should be the solution to Wikiquote? Maybe a simple interwiki redirect to SqWiki? Maybe even on SqWiki it can be a simple interwiki redirect to EnWiki? Maybe we can create a small explanatory page, a help page, that has 1 single paragraph for each entry and the links could work with anchors? (The last proposal is related to SqQuote.) - Klein Muçi (talk) 12:32, 18 October 2020 (UTC)

Also, I must say I copied the version you already have here without doing any changes. That means the redirects are also copied like they are here and that's another dilemma for me: Should they? - Klein Muçi (talk) 12:38, 18 October 2020 (UTC)

I hadn't thought of this before, but for the cases where sq.wiki does not have an identifier article, you can use the redirect to interwiki-link to en.wiki (or any other wiki that you choose). So, instead of:

redirect = 'ASIN (identifier)',

you might write:

redirect = ':en:ASIN (identifier)',

or even:

redirect = ':fr:Amazon Standard Identification Number',

Remember to include the leading colon.

If that is sufficient then creating all of those stubs is not necessary.

—Trappist the monk (talk) 14:09, 18 October 2020 (UTC)

Well, I just did that and it works (at least, it does with ISBN, that's what I tried). It worries me a tiny bit that by doing so, I've removed the possibility of having them in Albanian. But maybe if we start having the needed articles, I can switch back. Can you also explain me how the id limit works and why does it change frequently between updates? - Klein Muçi (talk) 14:37, 18 October 2020 (UTC)

The limits exist to catch simple typos: too many digits, most significant digits out of bounds. Alas, we can't catch too-few-digits or typos that produce in-bounds results... cs1|2 can't do much more to protect editors from these kinds of mistakes. The limit should be sufficiently tight that we catch typos but not so tight that we overrun the limit every few days.

I suspect that because of COVID-19, scholarly and academic authors are creating a flood of new papers so we are having to adjust the limits more often.

—Trappist the monk (talk) 14:51, 18 October 2020 (UTC)

Interesting... And finally, do the sandbox versions change substantially from the actual versions? Should I be on the lookout for any new upcoming update any time soon? (Maybe I should update now from them?) - Klein Muçi (talk) 15:12, 18 October 2020 (UTC)

I don't think that it is a good idea for you to update from our sandboxen. There is no guarantee that what is there today will be there tomorrow; no guarantee that what is in the sandboxen isn't broken.

There are no plans for an immanent update.

—Trappist the monk (talk) 15:26, 18 October 2020 (UTC)

Good. Okay then. That concludes it. See you on the next discussion. :) — Preceding unsigned comment added by Klein Muçi (talk • contribs) 15:35, 18 October 2020 (UTC)

I'm back sooner than expected. I was checking the categories (all of them) to see if I had missed something in the past updates. (I had missed only the ISSN one and the one regarding the test of location.) I was a bit confused with the display-name one but I think I fixed that one. (Can you take a look at the Albanian one and if it works fine with our module? Just to make sure.) But I'm really confused with these 2: Category:Pages with inactive DOIs and Category:CS1 maint: DOI inactive. Can you explain to me how those are supposed to work and what I should do with them? I guess I should create 2 separate categories. We already had the first one but I moved it to the second now. Only now I see they are supposed to stay together (no?) and I've messed it up. It was easier when 1 line of code = 1 category. These edge, new cases are making it a bit hard for me to follow through because of the complexity, even though I understand that increased complexity = increased efficiency in most of the time. - Klein Muçi (talk) 02:50, 19 October 2020 (UTC)

The several display-names maintenance categories have become a single display-names error category so your category name should reflect that. Our new name is Category:CS1 errors: display-names.

We are standardizing category names. The Category:Pages with inactive DOIs categories were kept in Category:CS1 maintenance‎ but weren't handled in the modules as maintenance categories and did not have maintenance category names. That has been fixed. The Category:Pages with inactive DOIs categories are being allowed to empty into the Category:CS1 maint: DOI inactive categories after which they shall be deleted.

—Trappist the monk (talk) 10:29, 19 October 2020 (UTC)

So we will have only 1 category, not both of them. Can you just help me and see if our category for display names works all right with our module? I have it linked in the English one, so you can find it easily.

Also, can some kind of test be done about categories in general? For example all categories that are used in the module appear somewhere (blue if they exist, red if they don't) how they are used in the module as soon as they are added there. That wouldn't help with the deleted ones but it would help with newly added ones and to check with name typos in the module. There are many cases where I've typed just 1 capital letter instead of a small one and the category wasn't working but I had no way to find out other than by chance because most of the error/maint categories are empty anyway in our wiki (and most of small Wikis). Imagine if one wiki decided just now to start using that module. They would have to accommodate an enormous number of categories to make it work which all need to be translated and have their description written. Most likely they would take some weeks, if not months, to fine-tune everything together in that case because of discrepancies that could be made between module and category names because of typos. - Klein Muçi (talk) 11:53, 19 October 2020 (UTC)

Umm, no you haven't. This is not correct:

category = 'Gabime CS1: display-$1',

at en.wiki it is:

category = 'CS1 errors: display-names',

I can probably write a module that will list the categories that are named in ~/Configuration; dynamically named cats will be problematic as are properties cats. Ping me if I forget to do this.

—Trappist the monk (talk) 12:39, 19 October 2020 (UTC)

So it should be Gabime CS1: display-names. Got that. They're not subcategorized further than that? I should delete the subcategories then. Also, I checked ALL the categories in EnWiki and SqWiki related to CS1. I found 3 of them which exist in SqWiki but aren't part anymore of the CS1 category in EnWiki: Category:Articles with missing Cite arXiv inputs, Category:Pages containing links to subscription-only content and Category:Pages with login required references or sources. What's the matter with them?

Regarding the module you say... I don't really know what you have in mind specifically but whatever that is, maybe it could help check even for deleted categories? By checking what the module currently has and if the "main categories" (error/maint/prop) have more than that, show them somewhere. Or if they have something which doesn't exist on the module because the name on the module is differently written (typos). So basically, we don't only check the module categories but we also check the "CS1" subcategories against the ones on the module and show the extra ones somewhere.

I'm almost sure I have A LOT of mistakes regarding categories now given that we have to keep up with the changes in 3 different ways with them. The SqWiki ones, the charts in SqWiki and the SqQuote ones, which change slightly from those at SqWiki. It would be good if whatever you mention above could be somehow integrated on the CS1 module itself so I wouldn't have to make place for a new module infrastructure, however small that might be. (On SqQuote new modules are a bit problematic.) But do as you see fit. Meanwhile I'll try rechecking manually all the categories between EnWiki/SqWiki/Charts/SqQuote/module in SqWiki/module in SqQuote to try and find discrepancies. :P - Klein Muçi (talk) 13:19, 19 October 2020 (UTC)

Category:Articles with missing Cite arXiv inputs is filled by {{cite arxiv}} for use by Citation bot which can fill a mostly-empty template. cs1|2 stopped using Category:Pages containing links to subscription-only content and Category:Pages with login required references or sources when |<param>-access= was created. These two categories are still used by {{subscription required}} and {{registration required}}.

I think that the best that the proposed module can do is make lists-of-links to categories. The lists can be separate: error, maint, some of the properties cats; perhaps with a heading that contains a count. Modules cannot look into existing categories and get the content so it is not possible for the module to compare what it thinks should exist with what actually exists.

I'm not interested in integrating this into the cs1|2 module suite; it's too big as it is. Might be made part of Module:cs1 documentation support.

—Trappist the monk (talk) 14:22, 19 October 2020 (UTC)

I was able to unify SqWiki and SqQuote in terms of modules and category names + the charts in SqWiki. I also compared the SqWiki categories against those at EnWiki and I think we have the same now (therefore the same at SqQuote too). The only thing that remains to be done now is to compare them against the ones we use on the module to check for typos. I will be waiting for your new module for that.

You are the master on that so I will take your word, whatever that is. But just for personal curiosity, what's the problem on having 10 subpages (or more) instead of 9? I don't think that there exists limits for that. Limits only exist on the length of the said subpages. Wouldn't it be easier to have everything related to 1 concept at the same place? For me personally it doesn't make any difference but if someone comes and wants to start from scratch, I think he would like to have things as tightly compacted as they can be, no? My idea here being why not make Module:cs1 documentation support part of the cs1|2 module suite itself? I guess as it is now, many wikis don't know about its existence. I'm just curious because, as I've said in the past, I care about the "meta point of view" in globalizing your work. - Klein Muçi (talk) 12:29, 20 October 2020 (UTC)

The purpose of the module suite is to render cs1|2 templates into readable citations. Module:cs1 documentation support contributes nothing to rendering cs1|2 templates so should not be integrated into the module suite. Still, at Module:Citation/CS1, the documentation module is listed so in that sense, everything related to 1 concept [is] at the same place.

—Trappist the monk (talk) 13:34, 20 October 2020 (UTC)

Oh, yes. I see it now. I've never stopped to examine all the elements in that page given that my aim when going there has always been in just copying the code. Well then, I'm waiting for your changes on that module then, whenever those come. Take your time but please, ping me when that time comes.

I saw below your invitation to take part on the discussions regarding WikiCite. The Albanian community has shown to me the WikiCite project a couple of times in the past too, given my interest with citations but I've found no substantial way to interact with it. If I'm not wrong, that's the second invitation you get regarding that subject. Maybe the correct way to globalize your work is through that initiative so I'm personally hopeful in your involvement. Maybe a more organic involvement with the Wikimedia Foundation could give you access to wider technical resources, liberating you from the technical limits that you're stumbling now with CS1 (and the i18n aspects of it). - Klein Muçi (talk) 17:18, 20 October 2020 (UTC)

Confused: I tried using nocat on a random article (take a look) for testing purposes. The category I get is in English even though it should be categorized on a different one in Albanian which already exists. What's happening? - Klein Muçi (talk) 13:19, 21 October 2020 (UTC)

After investigating a bit, I now see something "strange" is set in the mainpage of the module regarding that parameter... - Klein Muçi (talk) 13:23, 21 October 2020 (UTC)

Perhaps this is a case of do nothing. Given the evidence so far, |nocat= is little used here in cs1|2 templates so will go away at the next update. I hacked the module after our update when I discovered the now-obvious flaw in the previous code. You can't categorize |nocat= when |nocat= is set because |nocat= inhibits all categorization including categorizing itself. If you want, in sq:Moduli:Citation/CS1 you can change:

table.insert (render, utilities.make_wikilink ('Category:CS1 maint: nocat'));

to

table.insert (render, utilities.make_wikilink ('Category:Mirëmbajtja CS1: Është përdorur nocat'));

—Trappist the monk (talk) 16:09, 21 October 2020 (UTC)

Hah! Yes, that's what I was thinking to do but I wanted to be sure before. - Klein Muçi (talk) 16:34, 21 October 2020 (UTC)

I have created Module:Citation/CS1/doc/Category list which is populated by cat_lister() in Module:cs1 documentation support. Interestingly, it shows a couple of shortcomings at en.wiki...

Automatically compares live to sandbox when sandbox exists; else just makes lists from the live module.

—Trappist the monk (talk) 17:35, 21 October 2020 (UTC)

Well done! I was able to identify 3 typos and 3 missing categories (the ones you are also missing).The same is true for SqQuote so double those numbers. I'm guessing the categories that have dynamic parameters as part of the title will be forced to stay "forever non-existent", eh? This is such an easy tool to have at use. I'm glad I asked for something like this. - Klein Muçi (talk) 02:25, 22 October 2020 (UTC)

line numbers

two discussion, two sections—Trappist the monk (talk) 18:04, 17 October 2020 (UTC)

@Klein Muçi: Unfortunately modules for me in SqWiki don't show the same as in EnWiki with numbered lines and syntax highlighting. That's strange since that's not the case for me, logged in or out. Have you by any chance disabled the editing toolbar (sq:Speciale:Preferencat#mw-prefsection-editing → "Aktizoni redaktimin e zgjeruar të shiritit të mjeteve")? Nardog (talk) 17:36, 17 October 2020 (UTC)

@Nardog: No, I have that activated in my global settings. But all my settings are global. It's strange that I get different results between different projects. In EnWiki it works fine, in SqWiki it doesn't. - Klein Muçi (talk) 17:52, 17 October 2020 (UTC)

Does it not show up even when you're logged out (e.g. in incognito mode)? Is the "<>" toolbar icon at top left of the editor also absent? Nardog (talk) 17:57, 17 October 2020 (UTC)

@Nardog: Apparently it does show up when not logged in. But when logged in the whole toolbar is absent, not just the "<>" symbol. Does that mean it is a problem related to preferences? - Klein Muçi (talk) 18:00, 17 October 2020 (UTC)

@Klein Muçi: Likely so. A gadget or a user script might be interfering. On a module page, press F12, choose the Console tab, and refresh the page and see if there are any errors. Nardog (talk) 18:10, 17 October 2020 (UTC)

@Nardog: It shows 2 errors (and 10 warnings). Should I bring the errors here? - Klein Muçi (talk) 18:15, 17 October 2020 (UTC)

@Klein Muçi: Perhaps we shouldn't bother Trappist any more, let's go to your talk or mine (or wherever you like). Nardog (talk) 18:23, 17 October 2020 (UTC)

@Nardog: Right. I'm writing at your talk page the errors I see. Thank you! - Klein Muçi (talk) 18:29, 17 October 2020 (UTC)

Nomination for merging of Template:Sclass

Template:Sclass has been nominated for merging with Template:Sclass-. You are invited to comment on the discussion at the template's entry on the Templates for discussion page. Thank you. --Trialpears (talk) 13:58, 22 October 2020 (UTC)

MonkBot

Any reason why edits like this can't be done via MonkBot? Headbomb {t · c · p · b} 17:08, 22 October 2020 (UTC)

Not enough to warrant dragging the bot through the WP:BRFA process.

—Trappist the monk (talk) 17:12, 22 October 2020 (UTC)

I'd say you have several thousand reasons for it. Especially since those edits flood watchlists. MonkBot already has approval for such edits [e.g. Tasks 16/17, if you want more, it's nothing a speedy approval can't handle]. Headbomb {t · c · p · b} 17:23, 22 October 2020 (UTC)

The vast majority of what remains in Category:CS1 errors: deprecated parameters is left there at the request of Tom.Reding (see Wikipedia:Bots/Requests for approval/Monkbot 17). In my current awb list there are about 800 articles to do.

—Trappist the monk (talk) 19:18, 22 October 2020 (UTC)

Until the next deprecation, which will again result in thousands of articles in need of updates. You might as well bit the bullet now and have a solution that scales and doesn't flood watchlists. Headbomb {t · c · p · b} 22:05, 22 October 2020 (UTC)

Lang in lang check

Hey, could you add to Module:Lang a check to see if parameter |1= already contains a lang tag? So something like if text:find("lang=") then tracking_category end? --Gonnym (talk) 10:05, 23 October 2020 (UTC)

You mean the html attribute lang=? Why? |1= is the language code positional parameter so whatever is there that isn't an ietf language tag is an error so is categorized, right?

—Trappist the monk (talk) 10:12, 23 October 2020 (UTC)

I thought 1 was the positional parameter for the text; so whichever is the positional parameter for the text. Want to check if there are situations of wikt-lang being used there. This is also relevant to the Template:wt TfD. --Gonnym (talk) 10:20, 23 October 2020 (UTC)

{{lang|<ietf tag>|<text>}}

Does this search answer that question?

—Trappist the monk (talk) 10:33, 23 October 2020 (UTC)

Yes, but as a tracking category it's much more friendly which was why I asked. --Gonnym (talk) 10:50, 23 October 2020 (UTC)

CS1 categories

I was checking for a final time for discrepancies in CS1 categories between SqWiki and EnWiki. Shouldn't EnWiki also have Category:CS1 maint: extra text: contributors list given that it does have Category:CS1 maint: multiple names: contributors list and Category:CS1 maint: numeric names: contributors list?

Also, can these categories:

• CS1 maint: numeric names: authors list
• CS1 maint: numeric names: contributors list
• CS1 maint: numeric names: editors list
• CS1 maint: numeric names: interviewers list
• CS1 maint: numeric names: translators list

and these categories:

• CS1 maint: multiple names: authors list
• CS1 maint: multiple names: contributors list
• CS1 maint: multiple names: editors list
• CS1 maint: multiple names: interviewers list
• CS1 maint: multiple names: translators list

be grouped in 2 bigger categories like these categories:

• CS1 maint: extra text: authors list
• CS1 maint: extra text: contributors list [?]
• CS1 maint: extra text: editors list
• CS1 maint: extra text: interviewers list
• CS1 maint: extra text: translators list

are in Category:CS1 maint: extra text? That's how we have them in SqWiki. Standardizing them would make checking for errors easier (at least for SqWiki) and also help globally in WikiData. Besides, shouldn't they logically be like that? :/ - Klein Muçi (talk) 12:57, 22 October 2020 (UTC)

Category:CS1 maint: extra text exists for its own sake; it catches variations of ed in |edition= and p, pp, etc in |page= and |pages=. These can, and probably should, be converted to an error message/category. Until that happens, I am more inclined to move the subcategories, out of CS1 maint: extra text so that they are visible. Because CS1 maint: extra text is default collapsed, its subcategories are hidden which doesn't seem like much of a benefit to me.

Yeah, we need to create Category:CS1 maint: extra text: contributors list. since created...

—Trappist the monk (talk) 13:36, 22 October 2020 (UTC) 13:41, 22 October 2020 (UTC)

Yeah, putting them under extra text was my idea. I've come around to Maybe Shouldn't Have Done That. That said, I think it would be nice to separate extra text into extra text edition and extra text pages.--Izno (talk) 13:54, 22 October 2020 (UTC)

Are you sure? I created all of the CS1 maint: extra text: <names> list categories and those creations put those categories into Category:CS1 maint: extra text.

I would not object to separating extra text edition and extra text pages into two error categories – no sense in keeping them as maintenance categories.

—Trappist the monk (talk) 14:07, 22 October 2020 (UTC)

Huh. I could swear I had done it at some point, but I see the change I thought I made in neither contributions nor deleted contributions.

No comment on error/maintenance. I don't think I've seen something when a citation has those issues that didn't need fixing, so even on the slight chance there is a false positive it's probably not all that false. (And we will see yet more OCLC bad metadata from Parsoid.)

I thought of a maintenance category we might want to entertain: "volume"/"edition" (maybe chapter?) text in title. Especially books and encyclopedias. --Izno (talk) 14:34, 22 October 2020 (UTC)

Ah, I see your rationale! Well then we can separate them (and maybe call "the main" extra text category something more distinctive, if so needed - but I don't think that's the case here). But can we also talk about the idea of grouping CS1 categories? Don't you think there's an organic need to organize them in groups (more than just error/maint/prop) considering the numbers there are there now? In the past I've tried grouping together some of them related to format errors (JFM, MR) or maybe those about ignored errors (ISBN, ISSN, DOI) but I found no strong shared characteristics between them to justify the grouping, so maybe I understand your point of view. But still, in general I think it's better to create sub-levels and maybe we can even "change" existing errors/maint mess. to go towards that kind of thinking? Somehow it feels kinda strange to have categories like those I've mentioned above ungrouped. What do you think generally of this kind of logic I'm following? More as a discussion than as a suggestion. - Klein Muçi (talk) 14:40, 22 October 2020 (UTC)

Apparently I'm alone at that kind of thinking. :P - Klein Muçi (talk) 13:40, 23 October 2020 (UTC)

How you group categories at sq.wiki is entirely up to you. You can define which category an error or maintenance issue will populate simply by setting the category value in sq:Moduli:Citation/CS1/Configuration. The individual categories that are thus defined can be grouped by simply adding category links to the individual categories that you have defined. The category structure used at en.wiki serves our needs but may not serve yours. You are not constrained to use our category structure.

As you can see from Module:Citation/CS1/doc/Category list, Category:CS1 maint: extra text will go away to be replaced with Category:CS1 errors: extra text: edition and Category:CS1 errors: extra text: pages.

—Trappist the monk (talk) 14:02, 23 October 2020 (UTC)

Yes, I know that. That's what I've been doing these times with the categories I proposed you grouped here too. (And by creating 1 extra CS1 category dedicated to language ones.) But I've found out that makes it hard for me to keep track of the changes with them and generally I'm all for standardization and... I wanted to know what you think too. There are times where EnWiki differs greatly from SqWiki (considering SqWiki is still a small one) but on this aspect, I think we don't and that's why I was interested to hear your thoughts. Because I was surprised with this statement: ...Until that happens, I am more inclined to move the subcategories, out of CS1 maint: extra text so that they are visible. Because CS1 maint: extra text is default collapsed, its subcategories are hidden which doesn't seem like much of a benefit to me. and the further Izno's support on the next comment. My general idea is that it's better to have many subgroups of categories in many layers because it makes up for a more organized overall environment which makes working with them easier. Keeping that idea in mind, if I was the main maintainer of the module, I would try and make the errors/maint messages similar in some aspects so it could translate in a more easy subgroup process in categories. Starting with that idea in the root. But on the other hand, you guys here seem to have the opposite idea to that, wanting to make the errors as distinctive as possible and keeping the categories as "closer to the surface" as possible. And you say that that helps in work with them generally. Given that you are usually more experienced than us, I was eager to know more details on that. How the community here generally organizes the work with fixing the errors. Maybe I learn something new. (It's not the first time I've changed "my beliefs" in relation to CS1 concepts after hearing your rationale. And I believe in this case, I will anyway, because of the strive to keep them as uniformed as possible with EnWiki.) - Klein Muçi (talk) 14:44, 23 October 2020 (UTC)

So far (so far) we don't have so many error or maintenance categories that they can't all be displayed on one page so why not show them all on one page? It used to be that category pages listing categories that include categories, would show the included categories if you clicked the blue ►. For example, in Category:CS1 errors, we have Category:Pages with archiveurl citation errors‎ which has the blue ► along with annotation that says that there is one included category – though I can't find it. Click the blue ► and you get the helpful message: nothing found. Because of that, editors cannot know if there are actual subcategories attached to the category or if the blue ► is just misleading – I couldn't find a category in Category:Pages with archiveurl citation errors‎ so is there one? What category is it?. Because I know this about the blue ►, I have generally ignored category names with the blue ► since the MediaWiki 'fix' that broke the previous functionality. I am glad for a simple flat category structure.

—Trappist the monk (talk) 15:45, 23 October 2020 (UTC)

So far (so far) we don't have so many error or maintenance categories that they can't all be displayed on one page so why not show them all on one page? I can give an answer to that. Because they may be already enough (read: many) to scare new editors, wikignomes, if we are to use wiki jargon, deciding to dedicate themselves in solving CS1 problems. "Hiding" them in sub-layers is maybe effective at getting new users interested, in a marketing-kind-of approach. Of course the total opposite might also be true: New users getting scared by the anthill they find. (You might have noticed by now that I tend to tend too much on new users needs.) But given the uncertainty of the subject (if you have insight to share here, I'd be more than willing to hear it) and the fact that maybe it's not an important detail, especially for big wikis which don't usually suffer for volunteers AND my desire for global standardization regarding citations, I'll follow your chosen infrastructure for categories. - Klein Muçi (talk) 16:06, 23 October 2020 (UTC)

Lua error in Module:Harvc

Thought you might want to know about this error:

Lua error in Module:Harvc at line 359: attempt to index field 'name_list_style' (a nil value).

--User-duck (talk) 18:15, 24 October 2020 (UTC)

Fixed, I think, thanks.

—Trappist the monk (talk) 18:22, 24 October 2020 (UTC)

Automatic row numbers

Tracked in Phabricator
Task T42618

Tracked in Phabricator
Task T203293

Hello (Trappist the monk), can you have a look at my post at the helpdesk? As far as I can see you created the template this post is about (I am not sure about this). --Kallichore (talk) 21:33, 25 October 2020 (UTC)

Yep, I created it. It is a hack to get around Wikitable limitations. The truly proper fix is for MediaWiki to add automatic row enumeration support to wikitables. The {{row numbers}} documentation does say that the template is a stop-gap until MediaWiki creates that functionality.

—Trappist the monk (talk) 23:03, 25 October 2020 (UTC)

Question about lang_xx error not working

Hey, was wondering if you might know why ar-Arab correctly produces an error but en-Latn does not? I'm looking at Module:Language/data/iana suppressed scripts and I see "en" listed there.

• {{Lang-en|script=Latn|text=text}} -> [text] Error: {{Lang-xx}}: script: latn not supported for code: en (help)
• {{Lang-ar|script=Arab|text=text}} -> [text] Error: {{Lang-xx}}: script: arab not supported for code: ar (help)

--Gonnym (talk) 17:44, 27 October 2020 (UTC)

Fixed in Module:Lang/data

—Trappist the monk (talk) 18:12, 27 October 2020 (UTC)

Thanks! --Gonnym (talk) 18:17, 27 October 2020 (UTC)

Categories' description

Hey there! Maybe you can help by adding the description on these 2 categories: Category:CS1 maint: ignored DOI errors, Category:CS1 maint: JFM format? Those are the only ones related to CS1 without a description, if I'm not wrong. Then I would know what to better write in their Albanian homologues. :) - Klein Muçi (talk) 17:54, 26 October 2020 (UTC)

done

—Trappist the monk (talk) 16:55, 27 October 2020 (UTC)

Thank you! :)) - Klein Muçi (talk) 01:47, 28 October 2020 (UTC)

Csi: Miami: Season 4: Episode: Driven

if You saw (Episode Driven) recently, then Y didn’t Dumb Delko ask Hayden to describe Thief’s face to Sketch Artist?(2603:3006:500:7D00:DCE8:A693:1823:BA23 (talk) 18:19, 28 October 2020 (UTC)).

Help with /testcases

Hey, I'm trying to setup /testcases for the make_translit() function in Module:Lang and I'm not able to get some of them to work at test_12_transl() (in the /testcases). I'm not sure how to reach the "no_std" tests for which Module:Lang/data lists only 4 entries ("akk", "sem", "phnx" and "xsux") yet "Cyrl" somehow reached that part. Could you help adding the 3 missing testcases and do you have any idea why "Cyrl" reaches that section of the code when it shouldn't? --Gonnym (talk) 10:26, 31 October 2020 (UTC)

Are you sure?

{{transl|akk|text}} → text

<span title="Akkadian-language romanization"><i lang="akk-Latn">text</i></span>

{{transl|sem|std=|text}} → text

<span title="Semitic languages romanization"><i lang="sem-Latn">text</i></span>

{{transl|phnx|text}} → text

<span title="Semitic transliteration"><span>text</span></span>

{{transl|xsux|std=|text}} → text

<span title="Cuneiform transliteration"><span>text</span></span>

{{transl|Cyrl|text}} → text

<span title="Cyrillic-script transliteration"><span>text</span></span>

The giveaway is 'transliteration'. Because akk and sem are language codes without script, they get handled as generic romanization. I'm not sure why these are listed in in the no_std table other than because they were in the wikitext version of {{transl}}. Scripts phnx (ISO 15924 Phoenician) and xsux (ISO 15924 Sumero-Akkadian cuneiform) are overridden by the no_std table. Cyrl (ISO 15924 Cyrillic) is not overridden so doesn't use the no_std table.

So, what needs doing, I think, is that akk and sem should be removed from no_std table. I also wonder if 'transliteration' where it occurs in translit_title_table and in Module:Lang should be replaced with 'romanization'.

—Trappist the monk (talk) 11:39, 31 October 2020 (UTC)

Regarding the correctness of their usage, I'll leave that up to you, as I'm really not familiar with that yet. Regarding the issue though, what I'm trying to reach is this section:

if is_set (tscript) then
	table.insert (tout, table.concat ({language_name, '-script transliteration'}));	-- write a script tool tip
elseif is_set (code) then
	if not language_name:find ('languages') then					-- collective language names (plural 'languages' is part of the name)
		table.insert (tout, '-language')							-- skip this text (individual and macro languages only)
	end
		table.insert (tout, ' transliteration');						-- finish the tool tip
else
		able.insert (tout, ' transliteration');						-- generic tool tip (can we ever get here?)
end

I don't know what is needed to reach these 3 sections. The Cyrl code works but I don't understand why it reaches this and the two code related examples I've yet to reach. --Gonnym (talk) 11:48, 31 October 2020 (UTC)

Ok, it seems I misread the last part of the code, not sure how. So "xsux" and "phnx" work here. That's great! The above code though I still can't reach. --Gonnym (talk) 12:05, 31 October 2020 (UTC)

For my examples, make_translit is called with

make_translit (code, language_name, translit, std, tscript, style)

1. make_translit (akk, Akkadian, text, nil, nil, nil)
2. make_translit (sem, Semitic languages, text, nil, nil, nil)
3. make_translit (nil, Phoenician, text, nil, phnx, nil)
4. make_translit (nil, Sumero-Akkadian cuneiform, text, nil, xsux, nil)
5. make_translit (nil, Cyrillic, text, nil, cyrl, nil)

Examples 1 and 2 are handled at line 631. Examples 3 and 4 are handled at line 659. Example 5 is handled at 664 (second line of your quoted code snippet).

—Trappist the monk (talk) 12:22, 31 October 2020 (UTC)

Are these the same examples you posted above? If so, none of them produce "x-language transliteration" or "x languages transliteration" which is what I'm trying to achieve. --Gonnym (talk) 12:32, 31 October 2020 (UTC)

Yes. Right, none of them should. I thought that I showed where they were all handled, did I not?

Is this another case of finally pulling enough teeth to get you to ask the question that you really want answered? Do you know how frustrating that is? By not asking a clearly stated question, I spent I don't know how long, apparently answering the wrong question. To get to the if ... then ... end at lines 658–673 tscript must have a value. That means that the elseif ... then ... else ... end at lines 665–672 is unreachable. The path is line 663 to line 664 because tscript set, and then on to line 676.

—Trappist the monk (talk) 15:03, 31 October 2020 (UTC)

I really appreciate all the help you gave me, but I really think the fault of not understanding lies with you here. While I did incorrectly state my original issue, I corrected myself here after you gave me your answer. That in itself was not wasted time either, as it gave me another pair of eyes and cleared up that section of the code to me. The second part, which was my code snippet above (taken directly from the live code) is very clear (or is clear as it is in the code itself). You choose to disregard that code completely and repeat the examples you gave in your first response. I always ask exactly the question I want answer. I might have mistakes in that question, but I never try and hide something. Why would I? Waiting for someone else to answer is very time consuming and wasteful. Regardless though, I find it quite unfriendly complaining that my questions have wasted your time, when that fact is that I wouldn't need to ask these questions if the /testcases were setup, instead of me trying to figure out other people's intentions in the code and how to make them work. Which then turns out that I've wasted hours myself trying to get something to work which you then say That means that the elseif ... then ... else ... end at lines 665–672 is unreachable. --Gonnym (talk) 15:18, 31 October 2020 (UTC)

Question

Hello Ttm, I am wondering if you know of a bot (or a script, or... whatever) that can identify text and automatically update it. For example, if there was a list of ships, written out in markup, eg; ''[[HMS Bounty]]'', could the list entries be automatically re-written to ship templates, eg; {{HMS|Bounty|1787|6}} ...? Any info would be much appreciated. Cheers - wolf 12:47, 29 October 2020 (UTC)

WP:AWB can do this. For your Bounty example:

1. make a list from a Wiki search (text) using the search string insource:"Bounty" insource:/\[\[HMS Bounty\]\]/ [results]
2. do a normal settings find and replace for wikilinks inside italic markup:

   f: ''[[HMS Bounty]]''

   r: {{HMS|Bounty|1787|6}}

3. then for wikilinks that are not inside italic markup:

   f: [[HMS Bounty]]

   r: {{HMS|Bounty|1787|6}}

the order above is important

—Trappist the monk (talk) 13:19, 29 October 2020 (UTC)

Thank you, much appreciated. - wolf 02:04, 2 November 2020 (UTC)

MathWorld overhaul?

Thank you very much for CS1'ifying the {{Cite IUCN}} family of catastrophes. That is very satisfying to see. Do you have any plans to do something similar to {{MathWorld}} and/or other CS1 wrapper templates?   ~ Tom.Reding (talk ⋅dgaf)  17:15, 4 November 2020 (UTC)

Pretty sure that I did not know that the template exists. I usually convert templates to use Module:template wrapper when something else breaks. I recently did a bunch of encyclopedia wrapper cites because they were contributing to Category:CS1 errors: empty unknown parameters. Is there a need to convert {{MathWorld}}?

—Trappist the monk (talk) 18:26, 4 November 2020 (UTC)

I've seen several uses of {{MathWorld}} with multiple authors in |author=, and, what's worse, with Eric Weisstein as the second author (a quick intersection with Category:CS1 maint: multiple names: authors list yields a count of 68 as the upper limit of offending cases, ex. Claw-free graph & Mertens' theorems). He's hard-coded & wikilinked in the template as |author=[[Eric W. Weisstein|Weisstein, Eric W.]], so even to correctly apply |author-link= would require a confirmation of Weisstein in the |author1= spot, which starts the case for Luafication, let alone for multiple authors not starting with Weisstein. To do it properly, I suppose, Weisstein should be appended to the author list IIF he has not been previously mentioned, to retain the originally intended functionality of the template. Currently, |author=Mugan, Jonathan William; [[Eric W. Weisstein|Weisstein, Eric W.]] is the "correctest" work-around for {{MathWorld}}... Ugh.   ~ Tom.Reding (talk ⋅dgaf)  22:39, 4 November 2020 (UTC)

Apparently I was aware of it because I had begun to adapt it to Module:template wrapper. Done now.

—Trappist the monk (talk) 00:44, 5 November 2020 (UTC)

ISBN

Hi, need your help to fix isbn error on bnwiki. Usually 13-digit ISBNs beguns with '978' or '979'. However, ISBN issued from Bangladesh begun with 984. This cause problem on bnwiki when we cite any bengali book as module gives "invalid prefix" message. I believe following code is responsible for this. I would to like made an exception for 984 on bnwiki. How should i rewrite following code? could you please help?

		if id:match ('^97[89]%d*$') == nil then
			return return_result (false, cfg.err_msg_supl.prefix);				-- fail when ISBN-13 does not begin with 978 or 979

Thanks. --আফতাবুজ্জামান (talk) 16:12, 9 November 2020 (UTC)

Are you sure? 984 is the registration group element for Bangladesh and is the second group in 13-digit ISBNs (after the 978 prefix): 987-984........ 984 is the first group in 10-digit ISBNs (984.......).

According to https://www.isbn-international.org/content/what-isbn, there are only two ean prefixes: 978 and 979. You would think that if there were another ean prefix, ISBN International would have declared that on their web page.

—Trappist the monk (talk) 16:57, 9 November 2020 (UTC)

You are right. But i found many book with this type of format. For example: ISBN 9847016800481 or ISBN 9844613437. I don't know what kind of formula they use to generate ISBN. I think i will use |ignore-isbn-error= to fix these error for now. --আফতাবুজ্জামান (talk) 18:19, 9 November 2020 (UTC)

That 13-digit number is likely an EAN (see also https://isbn-information.com/the-ean-system.html) so even though it is printed in books as an ISBN, it is not. EANs and ISBNs use the same check-digit scheme so when the cs1|2 978/979 test is bypassed, the check-digit validates. Still not an ISBN because wrong prefix. The 10-digit ISBN seems to be a misprint. This link suggests some possibile reasons for the cs1|2 error: https://www.isbn-check.de/checkisbn.pl?isbn=9844613437&submit=test&lang=en; (isbn-check.de is blacklisted at en.wiki so I can't make it an actual link)

—Trappist the monk (talk) 19:30, 9 November 2020 (UTC)

Adding a new parameter to the CS1 templates

Hey Trappist, I was wondering if you would be willing to help out with a new parameter in the CS1 templates. We've had a significant issue just figuring out how to best handle linking book references to books. Then it hit, this could be a whole simpler if I could just throw an identifier into a parameter and let the cite template figure out how to best render what it has. Should be simple enough to install, shouldn't affect existing output, so should be uncontroversial to add, and eliminates the potential of creating unexpected errors down the road. Thoughts? If you are interested, I would like to work with you to get the technical details worked out.  :-)—CYBERPOWER (Chat) 19:17, 13 November 2020 (UTC)

Any additions to cs1|2 are best discussed at Help talk:Citation Style 1. What is it that you really want to do? I don't get a clear understanding of the problem from what you've written here.

—Trappist the monk (talk) 19:37, 13 November 2020 (UTC)

Trappist the monk, linking to books at archive.org. Many book references that have a URL go to something that doesn't actually provide meaningful information, just on where, or how, to acquire it. But some URLs are good links. It's hard to distinguish the two, so the idea is to supplement. We've established that we aren't replacing any links on references, but that leaves out a lot of references that can potentially be linked to because there's either a title-link in place, or a meaningless URL. —CYBERPOWER (Chat) 19:50, 13 November 2020 (UTC)

I think that you should make your case for a supplemental parameter at Help talk:Citation Style 1.

—Trappist the monk (talk) 20:15, 13 November 2020 (UTC)

Csi: Miami: Season 4: Episode Driven

If Posible reply Under each Question slowly:

1. Did you watch (Driven) recently?

2. Why didn’t Dumb Cop Delko ask Hayden to describe (Thief’s face) to (Sketch Artist)?(2601:204:D980:9640:586:7297:FBCF:54F3 (talk) 18:58, 15 November 2020 (UTC)).