Page MenuHomePhabricator

Special:NewPagesFeed duplicates entries when you scroll down
Open, Needs TriagePublicBUG REPORT

Description

Reported by @Fram at https://en.wikipedia.org/w/index.php?title=Wikipedia:Village_pump_(WMF)&diff=prev&oldid=1346855660

As for how to reproduce this, open the page and scroll down. My settings are "New Pages Patrol", active filters both reviewed and unreviewed, and both nominated for deletion or not. If I right now completely refresh the page, it starts with "Temper (Fargo novel" from 10:12 2 April, then "Kamelwa cave" from 09.23 3 April, and then it works apparently fine until "amorphochelus minutus" from 08.45 3 April. Scrolling down then triggers a load, and the next page I see is again "Temper (Fargo novel)". So in this case 20 pages OK, and then the doubles start to appear (the next page which appears is "Athalie Manvoy", which was the 3rd originally: no idea why the second one is skipped in the doubling). It seems to be that after each batch of 20 or so pages, the tool fetches more pages, but this includes a whole lot of doubles, triples, quadruples. The page also ends after 160 or pages (including the doubles), which is way too short.

Steps to replicate the issue (include links if applicable):

What happens?:

What should have happened instead?:

Software version (on Special:Version page; skip for WMF-hosted wikis like Wikipedia):

Other information (browser name/version, screenshots, etc.):

Event Timeline

Restricted Application added a subscriber: Aklapper. · View Herald Transcript

I was unable to reproduce. I used the settings you mentioned, then scrolled down until there were 320 entries, then I copy pasted the page into Microsoft Excel. Then I turned on Conditional Formatting -> Highlight Cell Rules -> Duplicate Values, which highlights duplicate cells red. Then I sorted A-Z and deleted any rows that weren't article titles or article descriptions to get rid of false positives. All 320 were unique.

Posted at https://en.wikipedia.org/wiki/Wikipedia_talk:New_pages_patrol/Reviewers#Special:NewPagesFeed_issues%3F to collect more information about whether this is common or uncommon.

Also suggested that a video be submitted to see if we can learn more steps to reproduce from that.

I found a second person who has this problem.

I've had the first issue even since I started reviewing for AfC some years ago. My theory is that if there's any sort of update in the list (e.g. brand new articles or articles being taken off the list) then all the entries just get repeated again from the top. I could be wrong. --Rambley

Sounds like we need to do some work on the steps to reproduce. Rambley's hypothesis is worth trying out: this bug may only happen when a new article is added to the queue, which can take a couple minutes in production, or may never happen when localhost testing unless you intentionally do it.

Someone pointed out that the test script https://en.wikipedia.org/wiki/User:JSherman_(WMF)/T414892.js exists to check Special:NewPagesFeed for duplicates. You install it, then click something in your toolbox to turn it on. Probably a bit more efficient than me copy pasting things into Excel and turning on conditional formatting -> highlight duplicates, then sorting.

cc @jsn.sherman who has worked on this style of bug before.