Both https://en.wikipedia.org and https://en.wikipedia.org/ now have their own short URLs. The former generates an error message, but the server shortened it anyway.
I think the extension should automatically add a slash to the end for such URLs. This would both resolve the issue and reduce the number of effective duplicates.
| Subject | Repo | Branch | Lines +/- | |
|---|---|---|---|---|
| Normalize plain domains to have a trailing slash after them | mediawiki/extensions/UrlShortener | master | +19 -5 |
| Status | Subtype | Assigned | Task | ||
|---|---|---|---|---|---|
| Resolved | None | T108602 Equivalent URLs are not canonicalized and deduplicated | |||
| Resolved | Legoktm | T220718 URLs with no slashes after domain name are "invalid" but are still shortened |
I just run into the same problem with https://de.wikipedia.org (= https://w.wiki/49) vs. https://de.wikipedia.org/ (= https://w.wiki/Ga).
I support the solution to enforce a slash on all URLs that don't have anything after the top-level domain.
I also run into a related issue in the process. For some reason I was able to submit https://de.wikipedia.org one time, but every later attempt to repeat that is now blocked with "Not a valid URL". I guess I was to fast and submitted the form without the JavaScript being fully loaded. And indeed, I can reproduce the issue when I disable JavaScript. So the next issue is: The frontend attempts to block URLs that are not blocked by the backend.
Change 512615 had a related patch set uploaded (by Legoktm; owner: Legoktm):
[mediawiki/extensions/UrlShortener@master] Normalize plain domains to have a trailing slash after them
When the patch is merged and deployed, we will have a surprising amount of entries in the database that is not accessible any more. This is not strictly a problem, just a bit sad.
I suggest to:
I'm not sure if this is worth it, or if it is even possible.
I suppose it's possible, we would need to write a script for it, test the script, and then run it. I don't think that's worth it, but I'm not going to stop anyone from working on it if they want to.
Change 512615 merged by jenkins-bot:
[mediawiki/extensions/UrlShortener@master] Normalize plain domains to have a trailing slash after them