In T207781#5193114, @Gopavasanth wrote:

I just tried out playing with the code, debugging a bit, I found the that highlightSameTagRevisions and highlightSameUserRevisions are almost using the similar code ( revs = this.revisionList.getRevisions() ) but this issues is not in the case of tags. I'm just wondering, Any insights please?

In T219727#5167641, @WMDE-Fisch wrote:

I also left a note to myself to check the logs, when this is starting to be live tomorrow, so we _really_ can be sure, the error does not appear anymore.

Tagged and released, can be deployed now @MoritzMuehlenhoff

Tagged and released

The .js or more specific the . in the title seems to be the reason for the popups showing in one case and not in the other.

In T222896#5170333, @MoritzMuehlenhoff wrote:

In T222896#5170319, @ori wrote:

I just came across https://www.mediawiki.org/wiki/Extension:Wikidiff2/Release_process. @Legoktm, who handles this nowadays?

@WMDE-Fisch did the last release.

The misbehavior with the link seems to be related to custom styling applied on the footnote links on en.beta.WP

I also left a note to myself to check the logs, when this is starting to be live tomorrow, so we _really_ can be sure, the error does not appear anymore.

/me asks for access for a sneak peak

In T220309#5155047, @Raymond wrote:

But I have another candidat to check: https://de.wikipedia.org/wiki/Datei:Bilet_RATT_1994.JPG . Two days ago I gave an OK for Commons but still I am unable to move it. Getting "The file cannot be imported, because it is not marked with one of the required licences.". Could this be a cache issue?

In T210281#5160031, @Lea_WMDE wrote:

If I'm not mistaken the default is to not have bolding on the backlinks (but en-wiki is not the only wiki who added it, we know at least of zh, ar and es who also bold them). In my opinion it would be great if wikis were changing to unbold characters, otherwise we would just be loading more and more styling on top of each other on the wiki pages.

Hmm I'm a bit unsure if the problem with the automatic bot imports is comparable to what the FileImporter does.

In T222441#5159391, @thiemowmde wrote:

We want to remove the old interface entirely, and are waiting for green light from our communication team as well as product manager. We will remove these tests as well then.

I would disable them only if they fail with every single patch. But this doesn't seem to be the case. Failures are rare and usually gone when re-triggering Jenkins.

These are the ruby selenium tests for the "old" UI. I think we are at a point were we also could just disable them because it seems very unlikely that we will keep that UI in the codebase. - Opinions @thiemowmde ?

In T220479#5157518, @Krinkle wrote:

Please convert to mw.loader.using() so that it doesn't become someone else's problem when the tests fail again if mediawiki.api isn't used by the log in page for some reason. This is a test for the rollback feature, not for the login page. And neither of these enqueues mediawiki.api. It is needed for an inline snippet, so should be ensured by that inline snippet.

In T219727#5143213, @WMDE-Fisch wrote:

Just from some first approach manual testing:

• I can provoke an edit conflict when creating a new page under the following conditions
  • I am not logged in ( when I am logged in - at least this method - does not work )
  • I click three times very fast the save button ( clicking very fast only twice does not seem to help )

Just from some first approach manual testing:

• I can provoke an edit conflict when creating a new page under the following conditions
  • I am not logged in ( when I am logged in - at least this method - does not work
  • I click three times very fast the save button ( clicking very fast only twice does not seem to help )

Can be tested on de.beta when creating references on the user page and changing the user's gender to female so Benutzerin: shows as namespace.

Can be tested on the testwiki on Ewiger Medaillenspiegel der Olympischen Spiele

The patch that implements the desired behavior is now in a state were it could be merged. But from the discussion on that patch I want to bring a valid concern back to the general discussion here:

WMDE-Fisch set the point value for this task to 5

In T220200#5129022, @awight wrote:

I couldn't find a CSS-only fix, but maybe we can implement in JS by pushing the gradient mask up by the scrollbar height?