To my understanding, big part of the issue with this is old contributions in the databases that can throw up if reassigned.

More in scope for the community tech wishlist than the developer wishlist, I think?

Removing wishlist tag. This is important but not something that would make development easier.

@Tgr Ok, I've added this as parent task of the mentioned above. Feel free to triage/close as duplicate/decline as appropriate. Regards.

Please note that while we've discarded global user merge for now, this Special:UserMerge would be useful, still, in some cases where stewards are required to fix SUL issues.

A couple of days ago being able to at least merge two accounts on mediawikiwiki would've been tad useful (request) 😿

The main problem with this is the large number of database updates that need to be made here, as well as lack of indexes on those tables (e.g. T105398). When we tried global user merge on Yuvi's account, it ended up doing a full table scan on the commonswiki image table and caused a decent amount of replag. At that time Aaron and I had brainstormed some ways to work around the lack of indexes by just going through the table in batches but I don't remember why it was never implemented...probably lack of time/resources.

What I said in T177410#3659562 is still true, and has probably worsened.

@Legoktm Seeing the issues you describe, and taking into count that the extension is unusable in our environment as it stands now; is there any reason why we still keep UserMerge on Wikimedia Production? I still think UM would be fantastic and useful, but if it is broken and cannot be fixed I'd say we should undeploy the thing from prod. I do not have all the data nor the knowledge to make such decision so I'm not opening a task. I'm saying all this with sadness, for the record. Thanks.