Also for suggestions for how to fix. The url:s in general aren't broken (ie. for example, if you have paid access to hs.fi then urls will work) and even if they would be broken it doesn't change it that currently conversion will lose original information. Sure this is user error in the end, but currently citoid implementation will guide user to make the error.

@Esanders If I understand correctly that there is already detection if reference contains template (example: <ref>{{url|https://www.google.com}}</ref>) and if template is detected then it doesn't show convert-button. How detecting non-bare url would be more complicated?

I made bug ticket T333521 for cases like "<ref>[http://news.example.com/title-of-article Title of Article]</ref>". Currently, the visual editor will lose information as it replaces the old reference with a result from citoid. It would need a review dialog for manually copy-pasting information from old reference to new reference or something.

As I wrote @UOzurumba in fiwiki's discussion the couple days of discussion time is not enough for community discussions. However, the problems what Kyykaarme described in 2021 are still valid.

Is there some documentation on what is WikiGPT and how it is technically implemented?

closing as it seems to work in all places where I tested it.

Updatescript likely was blocked by longer running process. However, I tested manually and changed crontab configuration and seems to work now.

Seems to work by quick testing through settings.

Made an inquiries and sent email to person who have originally stored least some of the photos.

@Aklapper Pushing this up again. Who/where should I ask to get this forward?

In T706#8462887, @Zache wrote:

Hi, I Am Zache (Kimmo Virtanen) and work at Wikimedia Finland. I would like to get subproject creation right for project WMFI which is Wikimedia Finland's umbrella project in Phabricator.

Hi, I Am Zache (Kimmo Virtanen) and work at Wikimedia Finland. I would like to get subproject creation right for project WMFI which is Wikimedia Finland's umbrella project in Phabricator.

What's the difference between syncedPages-* and reviewedPages-*?

In T323495#8415316, @Tacsipacsi wrote:

Having up-to-date data is important, especially for motivation. If I worked hard to get a counter in the namespaces section to zero, and it’s still non-zero after hours, it’s disappointing. A better solution would be still collecting the data frequently, but dropping some older rows (e.g. the previous one unless it’s the one kept as daily statistics).

Here is some analysis.

• Columns reviewLag-anon-average, reviewLag-user-average, have meaningful data in some wikis (such as trwiki and ukwiki)
• data in other fields is pretty much repeating values

! In T323495#8419025, @Ladsgroup wrote:
Do we need all of these? Which one is safe to get rid of? Each type removed translates to millions of rows being cleaned up. Maybe syncedPages-NS*?

@Multichill or somebody, if you have time to check this before tomorrow (I have an IRL meeting with the students) it would be nice.

Yeah, and i have explained here why the data for getting trends is important T44360#3235893.

And it would make the statistics gathering multiple times harder at same time. Also, even if there is millions rows the amount of the data in the disk is trivial.

@Ladsgroup I have used the data for generating the crosswiki reports on how flaggedrevs is working so even it is kind of broken it works for that. Also some data works on somewikis and not on others. No idea why.

There is some older tickets related to this: T163107, T44360 , T177951.

This got my heart going. The current server of the map warper uses Bullseye (I'm the maintainer of only that instance in the project)

@Susannaanas & all, if I understand correctly, the current virtual server of maps-warper uses Debian Stretch as an operating system whose end-of-support is in 2022, and it should be updated to a new version of Debian. (T306080)

If understand correctly this is for shutting down https://warper.wmflabs.org ? It seems that it has least some active users.

I am still a little bit puzzled about what this needs to get forward. Is it a community process or WMF search team actions?

If it is running as an separate service and data would be queried via federation then SPARQL-anything could be used

• https://github.com/SPARQL-Anything/sparql.anything

Also note there is temporarily workaround in fiwiki by me using abusefilter to disallow hs.fi link edits by InternetArchiveBot.

About the protocol column. Currently to find the specific link you need to search both http and https prefix and moving the protocol out from the index column it would simplify the querying and likely makes it faster.

Risukarhi's description of how this happens was that.

1. Risukarhi reverts edit(s) using undo
2. In editbox Risukarhi edited or partially restored the unreviewed edit