Hi @Jdforrester-WMF ! Does this issue still exist? If so, could you add the steps on how to reproduce this behavior and what the desired outcome should look like? TY

Hello @gengh || @Jdforrester-WMF

• How do these files differ from the purpose of data/errors/error_types.yaml?
• When we consolidate, could we store said errors we need from error.js into a yaml file? So that using them in both JS and PHP is easy?

Hello! Is this still an issue? Also, what do we mean by 'ULS'?

Hello! Does this task still exist as an issue or are there any updates to this?

To confirm, this is still occurring and we still want to make this change when in 'Read' mode bc it is fine in 'Edit' mode?

I really like the idea of this functionality

Hello! Any updates on the discussion/thread? Have we decided on what to go with?

I notice the last thread here was a couple years back haha, is this still blocked? And does the checklist need updating (aka. is this still an issue or some part(s) of it)?

I would be confused having random lang aliases shown to me. How does the 'system' know what languages 'are related to me'? Does this mean, if in the case I have logged in in English, Hebrew, and Arabic at some point?

To clarify, the acceptance criteria is to see "Enter Text" in the respective user language and not in English for every language?

• Do we expect to have translation for every language or can we have a fallback language with a chip?

Update:

• regardless of 'successful' result from Beta or Prod, we don't actually know for sure if Orch is truly routing the Z61 refs appropriately bc we are mocking this behavior kind of.
• in order to know for sure, we need testing in the orchestrator

Hello! What is the best way to know/verify this:

if tests had undetected failures, they should be fixed

(Un-assigning myself for now as I'll prioritize other 'fix-it' tasks; am happy to pick up later or not, TY)

Current status: we think we've seen this configuration/routing fail either or both in Beta and Prod so we need to verify which/where and why, then how.

The checklist/task will be tackled in separate patches

Observations so far:

• _mocha test/benchmarks/javascriptEvaluatorBenchmarks.js takes forever (10 minutes in, it's still going⏳)
• _mocha test/benchmarks/pythonEvaluatorBenchmarks.js immediately returns:

(node:48369) [DEP0040] DeprecationWarning: The `punycode` module is deprecated. Please use a userland alternative instead.
(Use `node --trace-deprecation ...` to show where the warning was created)

If the status on this is still the same, would it make sense to mark this as 'Low'? Sry to intrude, just scouring for a good BE task to pick up!

Ah I see~ @cmassaro
Also, until Geno comes back:/resources/ext.wikilambda.edit/components/default-view-types/ZObjectStringRenderer.vue

That error looks familiar @cmassaro 💀

Noting I'm having trouble reproducing this issue; below is brief of what I tried:

To add to this; a specific example of why this will be helpful can be seen when we are testing out an implementation of a function. E.g. https://www.wikifunctions.org/view/en/Z13453, we click on 'details', we do not see which implementations were run to gather test results; when for this function there is both a JS and Py implementation. As a user, I would want to know which implementation(s) was used

Hello! I have tried out this using the provided link, using a string reference as well as string as input and it seems this is no longer an issue. Will move this to 'Verify In Prod' for now~

From my end, since docker compose exec mediawiki php maintenance/run.php update was giving me the following (stopping at random Z objs) several times:

To clarify following the discussion thread, we want to:

• show an empty string, "" if none has already been selected (instead of "Select function")
• but if said function has already been selected, we want to show that (instead of "Select function")

Hello, the link mentioned a couple times here, shttps://www.wikifunctions.org/view/en/Z10601 just send/redirects me to https://www.wikifunctions.org/wiki/Wikifunctions:Main_Page 🤔?

Hi @Jdforrester-WMF! I am not able to reproduce this in my local environment atm.

1. I made some revisions to a created function w/ implementations
2. Clicked on the revision history to view a single page of an old revision (w/ 'oldid')
3. Clicked on 'edit' and the function and persistent obj fields have not changed

I think the nit should be a new task because it seems a bit more complicated than just adding an isolated way to read LTR or RTL just from the text box. If I go to https://www.wikifunctions.org/view/he/Z10000 (already set to a different language, for ex. Hebrew), the page knows to set RTL; however, if we are wanting just the text box to detect a Hebrew selection when we're on an English page (i.e. https://www.wikifunctions.org/view/en/Z10000), manipulating that does not seem straightforward to me. What do you guys think @AAlhazwani-WMF @Jdforrester-WMF @gengh

In T353236#9567209, @AAlhazwani-WMF wrote:

nit: if we could display the char counter on the right (for LTR languages, and left for RLT) it would be great! similarly to this demo https://doc.wikimedia.org/codex/latest/components/demos/field.html#fieldset-with-custom-help-text-content

Ok so it was RTL, thank you 😸!

Hi @AAlhazwani-WMF what do we mean by 'LTR' and 'RLT'? Google says, 'long term relationship' and 'revocable living trust' 😆 .

In the code (WikiLambda/resources/ext.wikilambda.edit/components/function/viewer/FunctionViewerDetails.vue) where this test data would get initialized, I see mention of this task commented, 'TODO (T310003): fetch these from a cached table in mediawiki'. Perhaps this is not a bug, rather a non-trivial task and we are intentionally using '-' for the time being 🤔 ? @Jdforrester-WMF do you recall/have thoughts?

Do we know when we began to see this bug? Around the date this was reported, late September 2023?

lol the description on behaviors love it XD

kept some notes in this doc

Nice!