Thank you for your quick answer, which seems perfectly logical. The site does not comply with the standards.
If there's no existing "don't change url blacklist" in which the domain can be added, don't spend more time. I've written a little python script to correct these occurrences daily, this way the impact is limited for French wikipedia.

It may also be easier to send email from on-wiki Special:EmailUser (API) using a dedicated account for this usage.

Looks fixed for me, thanks @Sfaci and @BTullis for the quick fix.

I don't know if it's just a temporary processing delay, or a breakage. But given the different user reports the same day, I prefer to fill a task.
Edit: definitively not normal, few frwiki bots are broken because of that.

I've seen several user feedbacks indicating that the quality of Deepl is much more advanced than other engines offer, to the point of not using the Content translation tool only because Deepl is not available in it.
I saw someone who developed its own wikipedia translator tool around Deepl API, indicating that "google's and Microsoft's translators are years behind Deepl".

Old task requesting ability to create custom Echo notifications: T58362: Allow users to create custom notifications onwiki.

Side note: French wikipedia community would like to have these settings not shown on articles (ns0). The addition of INDEX and similar magicwords is currently blocked on this wiki by an abuse filter. Majority of users don't understand why the whole edit is blocked because they hit an obscure button that is actionable, but that they should not touch :)

got that too on frwiki, in my case error 500 while loading DiscussionTools "Response" form during few minutes, VE preview API calls were failing.
there was a spike of error 500 from around 21:54 UTC to 21:58 https://grafana.wikimedia.org/d/000000503/varnish-http-errors?orgId=1&from=1705604739966&to=1705615539966&viewPanel=7
and a drop of half ot the edits https://grafana.wikimedia.org/d/000000208/edit-count?orgId=1&from=1705613802299&to=1705615602299&viewPanel=8

I was following https://wikitech.wikimedia.org/wiki/Help:Toolforge/Raw_Kubernetes_jobs#Virtualenv_and_pywikibot on toolforge.
mwparserfromhell was missing from the default packages present on the system, and a nice error message asks to install it. But if user is using OAuth, pywikibot does not suggest installing mwoauth / requests_oauthlib.
We can also update the doc to ask running pip install -r /data/project/shared/pywikibot/core/requirements.txt.

Is it also substitution of {{reflist}} template, similar to what we for citations in T213277: CX2: Should not substitute citation templates when publishing ?

I guess this is the visible part of this main issue: T213277: CX2: Should not substitute citation templates when publishing

This bug is really the worst of the ones opened on ContentTranslation in my opinion, given the time spent by all the volunteer contributors who have to go back behind a translation to try and reformat the translations that have been substituted. It literally freezes the citation in a version that can't be edited later, except if somebody spend some time to fill a template. Not good to have such non-editable stuff in our wikis !

Still seen daily in 2023: https://fr.wikipedia.org/w/index.php?title=Utilisateur:G%C3%A9rard/Dolores_Donlon&diff=prev&oldid=210406796
https://sigma.toolforge.org/summary.py?name=Framabot&search=%2C+%7B%7Br%C3%A9f%C3%A9rences%7D%7D%2C&max=500&server=frwiki&ns=&enddate=&startdate=

Pull request: https://github.com/huggle/huggle3-qt-lx/pull/364

Thanks @Mvolz !

Thanks for the quick support :)

i guess i just deleted "~/.kube/config" by mistake last time... Could someone help me to get it back again? Thanks!

I don't know what sites table contains, it's not managed by Quarry that just shows the content of the replicas db. See doc on mediawiki.org.
Another way to list wmf sites is using meta db, this one contains correctly wikifunctions: https://quarry.wmcloud.org/query/75687

now functional: https://quarry.wmcloud.org/query/75686
According to the code, the dropdown menu just lists all databases that have already been queried by at least an user on quarry, original way to do caching but it works :)

Can I suggest following change: https://github.com/toolforge/quarry/pull/23. Keep message really visible (goal was initially to show a temporary important message). But reduced its size according to your suggestion, and contrast improved too.

Is it planned to fill a page on https://wikitech.wikimedia.org/wiki/Incident_status ?

take a look on https://wikitech.wikimedia.org/wiki/Password_and_2FA_reset#For_users

frwiki user reported that the tool didn't list his edits with summary Wikification (#moiswikif). Adding this char to the regex would probably be acceptable.

This is still happening daily with Author field containing facebook link, Ie https://fr.wikipedia.org/w/index.php?title=Lekima_Tagitagivalu&diff=prev&oldid=205847137 because <meta property="article:author" content="https://www.facebook.com/ActuRugby">.
I've made hundred of edits to remove these errors on frwiki: https://fr.wikipedia.org/w/index.php?title=Sp%C3%A9cial:Contributions/Framawiki&target=Framawiki&offset=20230706113753
At least Citoid should avoid using fields that are url.

sorry for my English. Instead of "one of two reports", I meant "a certain number of the changes reported". I didn't do the math, but it's half or a quarter of small changes to ignore. Examples: 1, 2 need to be closed manually because they're two small.

In T302154#8914904, @Andrew wrote:

Ah, sorry, I should've read back further in the task! Yes, that host can+should be deleted.

It also impact the ability to use the OpenRefine tool that is integrated inside PAWS: it doesn't handle OAuth, so wiki credentials need to be entered to edit wikidata from it.

I didn't want to propose a commit, it's risky to break the threading by putting a set() above the generator.

I have cleanup few items using this request: https://w.wiki/6nRN

In T169452#8886544, @IKhitron wrote:

Well, as a regular and very pleased user of Quarry, I tried Superset. Not helpful, it doesn't have most of databases. So, Quarry doesn't work any more, and I can't use Superset, so there is no way for me to run an urgent query (72767).

https://github.com/toolforge/quarry/pull/22/

In T302154#7724373, @Stashbot wrote:

Mentioned in SAL (#wikimedia-cloud) [2022-02-20T19:49:50Z] < @Andrew > moving nfs service from quarry-nfs-1 (bullseye) to quarry-nfs-2 (buster), testing to see if T302154 is a kernal or nfs-version issue

Add API endpoint(s) to https://hashtags.wmcloud.org/ that can be used to replace the prior database queries.

I've tried Superset for the first time and found some small points:

• query csv exports have a b' prefix in each cell. Here is the export of gendergap query:

https://github.com/toolforge/quarry/pull/20

@zhuyifei1999 any usecase in mind that need life-storage of the results? it may be a possibility to delete results after some time (ie. after 2 years). After such time links to queries shared on wikis are far less relevant.

To understand correctly, is this docker image used anywhere yet? I guess server is still using git clone and virtualenv without any container stuff ?