In T86910#11397002, @Samoasambia wrote:

@Pppery I noted that there are two spaces between "[[:$1]]" and "by" in the new message.

In T86910#11396424, @Pppery wrote:

A "dummy revision" is https://www.mediawiki.org/wiki/Manual:Dummy_revision

The new message says: because {{GENDER:$2|he|she|they}} made only dummy revisions.

In T402297#11384630, @Amire80 wrote:

This adds the message diff-usergroups. Currently, its English text is "$1 user groups" and its qqq is "The text for displaying user groups on diff view."

$1 is not documented in qqq. As far as I can see in the code, it's a number of user groups.

Is this number always more than 1? If it can be 1 or 0, the English message text should probably be something like "$1 user {{PLURAL:$1|group|groups}}". Even if it is always more than one, it should use {{PLURAL}} because many languages use different forms for 2, 4, 6, etc.

In T407653#11391416, @Niharika wrote:

In T407653#11352487, @Niharika wrote:

My proposal is to simplify the message language to avoid the "last" terminology -

Account creation is temporarily restricted because visitors from this IP address have exceeded the number of account creations currently allowed. Please try again later.
Kindly note that there is a limit of one account creation allowed every 10 minutes with a maximum of 6 account creations permitted every 24 hours.

This informs them about the limits but doesn't tell them which of the two limits have been tripped. I think it's OK to do this assuming good faith editors will likely try again later and vandals will just move on.

@Amire80 thoughts on this proposal?

This adds the message diff-usergroups. Currently, its English text is "$1 user groups" and its qqq is "The text for displaying user groups on diff view."

In T407580#11341770, @Michael wrote:

Thank you! Useful to know about {{experimental}}! Are these templates documented somewhere? I glanced at https://www.mediawiki.org/wiki/Help:System_message#Message_documentation but did not find guidance about that, but maybe I missed it.

Thanks for asking for my opinion :)

I think that when a message is marked as ignored, translations are supposed to be deleted from the JSON files automatically, but @Raymond should confirm.

In T396889#11304182, @Tacsipacsi wrote:

English Wikipedia does this: https://en.wikipedia.org/wiki/MediaWiki_talk:Accesskey-pt-watchlist (Template:Interface explanation). I vaguely recall having seen this elsewhere as well, but I’m not sure where (if at all).

In T396889#11223823, @Tacsipacsi wrote:

Sorry for being a bit late for the party, but there is one problem with this change, which should be solved one way or the other: wikis often link to translatewiki.net for message documentation, and now these documentation pages are abandoned and won’t be updated anymore. Ideas:

• Obvious: revert the change. This isn’t necessarily the best idea, though.
• Create a new message group called “MediaWiki – ignored”, which behaves much like a regular message group, but all its messages are automatically ignored, still saving the maintenance burden.
• Display the message documentation in a different way. Note that message documentation often contains translatewiki.net-specific templates, so it needs to be parsed on translatewiki.net, no matter where and how it’s shown. (And preferably stored in a way that backlinks show up on translatewiki.net.)

Hi. Some people expressed objections to this idea. It's my fault—I should have written more details earlier. I've updated the task description now. I'll reply in detail here:

In T407507#11287961, @Bugreporter2 wrote:

In T407507#11283866, @Amire80 wrote:

The current spelling came from a community discussion here: https://ig.wikipedia.org/w/index.php?title=Wikipedia:Nz%E1%BB%8Dr_ogbako&oldid=171993#Namespace_names

If you read that discussion, you'll see that they wanted Ńkàtá, but something went wrong in the copying process and what should have been N-acute (U+0143) got replaced with N-dot (U+1E44).

It's probably not related to Growth, but to the Edit check feature of Visual editor. Adding those tags. Revert if I'm wrong.

Oh, oops, should have been more careful. I'll revert to quotation marks.

• If available, URL to get the prior revision segment
• If available, URL to get the following revision segment

In T407507#11287868, @KingChristLike wrote:

In T407507#11284629, @Amire80 wrote:

In T407507#11283989, @KingChristLike wrote:

Hi @Amire80, I confirm that the new proposed spellings are correct.

Can you please link to a new discussion about it in an Igbo wiki? The discussion itself can (and should) be in Igbo.

Here’s a new discussion about it.

In T407246#11286012, @kostajh wrote:

The acct_creation_throttle_hit message has existed in core with the $2 parameter for a long time, so there are many translations that make use of this parameter already.

Can you please clarify more about parameter $2? What kind of values can it have?

In T407507#11283989, @KingChristLike wrote:

Hi @Amire80, I confirm that the new proposed spellings are correct.

Another link that may be useful: Here's the change from the older namespace names to the current ones: https://gerrit.wikimedia.org/r/c/mediawiki/core/+/1049639

In T407507#11283805, @Bugreporter2 wrote:

It would be helpful to have input from someone familiar with the process of renaming namespaces due to spelling mistakes.

1. Should the incorrect name be retained as a namespace alias for backwards-compatibility, or should it be scrubbed for simplicity?

I've already fixed "cope" in https://gerrit.wikimedia.org/r/c/mediawiki/extensions/ContentTranslation/+/1196419, and it should appear in translatewiki some time soon.

As I wrote in https://github.com/wikimedia/jquery.i18n/pull/307 , French probably shouldn't be defined as fallback, unless there's a strong and specific reason for it.

I think it's mostly moot because the localization statistics of Galician are usually 100% or close to it, so fallback is rarely seen in practice :)

In T392749#11196158, @Winston_Sung wrote:

In T392749#11193458, @Amire80 wrote:

Can you please link to more community discussions about it, in any language?

Other than the English Wikipedia and this Phabricator task, is there an external source for the "Cou" autonym? For example, a dictionary or another website?

In T400578#11215972, @Agabi10 wrote:

I might be misunderstanding because I lack context, but if this is to make modules translatable wouldn't a slot on the Module page be better for that? Instead of having to add magic to the title to set the content model it would be tied to a specific slot of the current module page. It's more feasible avoiding conflicts with other things that way too.

I have some questions because this adds an unusually high number of new codes, and this is very often a recipe for mess.

In T404534#11190324, @Jdforrester-WMF wrote:

i18n clean-up is not a bug. :-)

In T404671#11190148, @Jdforrester-WMF wrote:

In T404671#11185458, @GrounderUK wrote:

To me, it looks more like the first “calls to” would be the preferable omission.

Yes — we changed the wording during development and clearly failed to remove the old bit!

What about “Pages with errors in function evaluations” and “Pages where system availability issues caused function evaluations to fail”?

For consistency, we're using Wikifunctions calls as much as possible, as most users won't know what a "function evaluation" is, but they might recognise "Wikifunctions" and go to the project chat for help/complaints.

This is not exactly what we call "lego". "Lego" is more like concatenating different parts of sentences under the assumption that the order of the words is the same in all languages. This is very bad and must never be done.

Can anyone please say more about the message codemirror-wikilint-unclosed? What can $1 be? Is it shown as characters or as words? If these are words, where do these words come from?

The text-transform: uppercase; rule on ::first-letter is not just unnecessary, but actively harmful, and not just in languages written in the Arabic script, but in all languages. Some languages may require a small letter according to their grammar, and forcing a capital letter would be wrong (such things sometimes happen in Irish and in some languages of South Africa and Taiwan, for example).

In T402697#11121285, @Bokareis wrote:

Does a new community discussion have to be put up for this task?

In T381577#11142147, @Bhsd wrote:

@Amire80 Thanks for your comments!

• $1 parameters are not documented. Please add explicit documentation for every $1 parameter in every new message.

What documentation do you mean here? qqq?

In T381577#11142973, @Bhsd wrote:

r1184397 cannot be automatically merged. Maybe I need to avoid XML-like tags?

Some comments about the translatable messages:

A comment about the "n quick" message (cx-sx-translation-suggestion-easy-sections): It appears to be shown in hard-coded parentheses. This is wrong for some languages, which use different parentheses characters. It's better to use the parentheses message from core, or to simply include the parentheses in the message itself.

About parsermigration-reportbug-feedback-title: how will it work in practice on Wikimedia sites? Will the reporting be on local wikis or on a central wiki?

In T401916#11087043, @DLynch wrote:

There's presumably languages where the most-natural translation would include switching the order of the reference and the message, thus the suggestion for $1, though it doesn't really work out to do that in this situation since it's a full block-level display of it.

In T401916#11086988, @Esanders wrote:

I think the issue is that the message looks like it is followed by another word, and in many languages that next word would affect the grammar of the translation.

In T357981#11068913, @Amire80 wrote:

• That |language=ruo (for explanation on what I mean, please see the above comments) results in "in Istro-Romanian" and not "in Istro Romanian"; Megleno-Romanian (ruq) has the hyphen displayed in this exact case.
• That Istro-Romanian (ruo) be available in Wikidata for me to sporadically add entries in the language

Can you please create a separate task for that?

In T357981#11067832, @Super_Dromaeosaurus wrote:

Hello, just to clarify because the comments above are long, to summarize these are the remaining requests:

• Remove ruq-Cyrl and merge ruq-Latn into ruq
• Replace native language name displayed by ruq from "vlăhește" to "Vlășească"/"vlășească" (whatever capitalization is consistent with Romanian (ro) and Aromanian (rup, or roa-rup), they are the sister languages and have same capitalization rules).

Perhaps it's because I tested changing a non-existent page's content model, as discussed at T400578. I guess it doesn't actually work, and it's not just weird, but really broken and shouldn't be done. And if someone does try to do it, it should probably be prevented.

In T400578#11043821, @Pppery wrote:

You can go to Special:ChangeContentModel and enter a nonexistent page, and then specify the Message Bundle content model to create it directly. Well, at least, you should be able to.

The most significant issue I encountered while preparing the examples for the video: T400578

In T379892#11020294, @abi_ wrote:

Pending:

• Get documentation reviewed: www.mediawiki.org/wiki/Help:Extension:Translate/Message_Bundles/Lua_reference - @abi_
• Prepare communication message - @UOzurumba
• Prepare video demonstrating the use of the library - @Amire80
• Update a small module on Meta-Wiki to use the new Lua Library - @Amire80

We're planning to reach out to the community for feedback before the end of this week.

I'm not sure that it's right to have it as a Phabricator task.