I hate this feature. It seems like almost every edit I make is reported as an edit conflict when it is not, as in this recent edit https://en.wikipedia.org/w/index.php?title=Sykes_Camp&diff=933743700&oldid=933743640. Unless this can be fixed, I hope this feature is not implemented.
Topic on Help talk:Paragraph-based Edit Conflict Interface
As far as we are aware of, there is no way the TwoColConflict feature can influence the number of conflicts users run into. All the feature is supposed to do is providing a better interface after a conflict happened. You can disable it in your "Beta" preferences and see if this changes anything. Are you still running into conflicts then? Please let us know!
I disabled the beta feature. It apparently wasn't the problem. I am still experiencing many false positive edit conflicts. How can I fix this?
We are very much curious and would love to know more about this. We created phab:T222805 to collect all we know about this issue. Unfortunately it looks like it is very hard to track. It might be related to the webbrowser or a browser extension you use. Possibly even some malware silently running on your computer. It could be your computer mouse or your hand accidentally doing double-clicks whenever you click a save button. If you have an idea, please let us know.
I can reproduce the issue very readily in my home laptop. I'll try disabling all browser extensions, experiment, and report back.
I'll add to this, I'm seeing this way, way too often. Craziest one, I was creating my 2019 talk page archive (https://en.wikipedia.org/wiki/User_talk:Ravensfire/Archive_11) and got the edit conflict on the page creation!
Did a very quick test on this scenario, where I created one archive (12) and quickly added text and saved, no problem. I created another one (https://en.wikipedia.org/wiki/User_talk:Ravensfire/Archive_13) where I clicked on Create Source, then ignored it, using other tabs and brower windows for several minutes, then added text and saved. Boom, edit conflict happened. I wonder if the time involved could be an issue, or using multiple tabs / windows or a combination of these?
If anyone encounters this repeatedly I would consider whether it might be a symptom of a mouse button beginning to flake-out. Junk edit conflicts can come up when multiple clicks register on the save button.
I think I recall a current-or-recent discussion at English Village Pump (Technical) where someone was getting edit conflicts and they eventually concluded it was a mouse issue.
It's not related to a mouse. It's browser-specific. It only happens with Chrome.
False positives happen frequently when I use Chrome and single-click the "Publish changes" button. (I ignore the conflict, because my edit went through.) When I double-click the button (even though a single-click is sufficient), I do not get the false positive. None of this happens with other browsers I use.
If anyone experiencing this bug is able to post a screenshot of the conflict, that might provide some helpful clues. The "tabs" and "double-click" theories are unlikely because it's impossible to create an edit conflict with yourself (actually, this has always been a dangerous feature but not something we can tackle at the moment). Maybe the browser is sending both a logged-in and a logged-out submission... More information needed!
See https://commons.wikimedia.org/wiki/File:Curb_Safe_Charmer_conflict_screenshot.jpg (diff#942741035)
Thanks! Well, it seems like you did get an edit conflict with yourself, which surprises me. You were making the second edit using the same username, I assume? Is there anything else you can say about what happened? From the edit summary, I'm guessing the first edit was made using a toollabs service "reFill 2"? I'm not familiar with that tool but according to the documentation you submit the page in a toollabs form, probably in a separate tab. This means the page you're working on will be updated relative to what is loaded in the other tab. Depending on how you arrived at the editing page, maybe it's targeting a specific oldid, or maybe you already have the editor open while you run reFill in the second tab?
Either way, now I can imagine why this is happening. Try reloading your editor tab after using reFill: #1 save any outstanding changes and arrive on the article's "Read" tab, #2 run Refill in the second tab, #3 reopen the editor. Hopefully this improves the situation for you!
@Adamw the reFill thing is a red herring. I've just done an undo on an IP editor's edit to the en article on Praveen Linga. I hadn't run reFill on this article - at least not recently. Up came the edit conflict comparison. I have a screenshot if you need it.
I can avoid an edit conflict if I click Submit twice, rapidly. But the editor reopens the article in edit mode.
Hi, thanks. These screenshots are interesting… The new interface might give a bit more information about the conflict, but even from this I suspect that something unusual is happening. It does seem that you were the only one editing, and from the timestamps it looks like you did successfully save a version of the article, immediately before seeing the conflict. I've created Phabricator task T246440 which might be related, I'm not sure yet.
I may have a better clue as to what the problem is. It seems when I click on the Submit button, whether with the mouse or by tapping a touchscreen, I may get the phantom edit conflict. However, if I instead press the enter button on the keyboard while the cursor is in the edit summary field (which I rarely did until now), there is no edit conflict.
Thanks a lot for this insight! As far as I know the only difference between the two actions is that clicking the button submits the buttons wpSubmit value, while pressing enter in a textfield does not. I had a quick look and could not spot any code doing anything with the extra value. That's puzzling. We will definitely look closer into this.
I confirm Spiffy sperry's experience, hitting ENTER does not cause a conflicting edit.
I received one false-positive edit conflict. Here's the diff. The previous edit was two months ago. I made an edit, set the edit summary, and I clicked PUBLISH. The new edit conflict screen came up. It gave me the option to merge my edits into the current version of the page, but I saw no conflicting content to resolve. I was puzzled so I checked the page history. I found my edit had already been saved, with no other recent edit I could have conflicted with. I then abandoned the unneeded edit-conflict screen.
Edit: I just had another one. Here's the second diff. The new event report would be otherwise identical to my report above.
I found no logging from these empty conflicts, which is strange. There should have been no way to show a conflict page without also logging... A screenshot of the conflict page might have additional clues, please snap one if this happens again.