This is caused by ia-upload adding pages that are marked as <addToAccessFormats>false</addToAccessFormats> in the scandata XML file, as mentioned by InductiveLoad in this comment.

Thanks for the quick fix. 👍 Unfortunately, AWB now requires users to be added to Project:AutoWikiBrowser/CheckPageJSON, even if that page does not exist on the wiki (e.g. en.wikisource). If the page didn't exist, then in the previous version of AWB, you could log in and edit. @Reedy Is this change intentional?

I'm now getting the "This user doesn't have enough privileges to make automatic edits on this wiki" error with all Wikipedias, even en Wikipedia.

Reposting from the discussion on Scriptorium: I'm still getting the same error here on enws and all other sites I've checked, except Wikipedia. With Wikipedias:

• If I try to log in to en Wikipedia where I'm added to Project:AutoWikiBrowser/CheckPageJSON, it works fine
• If I try to log in to one which doesn't have Project:AutoWikiBrowser/CheckPageJSON (the first one I found was be-tarask Wikipedia), it logs in fine
• If I try to log in to one which has Project:AutoWikiBrowser/CheckPageJSON, but I'm not added to it, I get a different error message "Yodin is not enabled to use this", and when I click OK, I'm taken to the relevant Project:AutoWikiBrowser/CheckPageJSON page

Is it possible that the rollback is delayed for some reason on non Wikipedia sites (caching, etc.?)

I've tried several times to upload File:The New Monthly Magazine - Volume 011.djvu, and had this error most times. For the current version of the file it seemed to reassemble the upload chunks and publish it successfully, but the resulting file seems corrupted: attempting to download it gives a 404 error, it can't render page images, but does seem to be able to extract the text layer (e.g. for pages on Wikisource).

Wikipedia:Manual of Style/Accessibility suggests using Template:TOC limit in some cases, but this still seems to be broken for Vector 2022. I was wondering if there was any progress on this? (T333017 doesn't mention any details, including whether it will address the TOC limit issue, and it also looks like it might have stalled?)

Working perfectly thanks! Much appreciated! 👍

In T341742#9011665, @Aklapper wrote:

Hi, this misses steps to reproduce. How to have/get some "page scan image"? Is this about Wikimedia OCR maybe?

As a temporary fix, I've found that whenever this happens, I've been able to refresh the page, and it will load the search results from the page URL.

P.S. For what it's worth, the email notifications I'm getting from Phabricator about this are in html, and this site includes html (default) and plain text email format options. Would it be possible to use this codebase, rather than re-inventing the wheel?