Character list

@Lea Lacroix (WMDE) and Ladsgroup: I've noticed that https://w.wiki/2w and https://w.wiki/22222222222w are both functionally identical to https://w.wiki/w. Was this intentional? Should this be mentioned in the documentation? Jc86035 (talk) 09:29, 14 April 2019 (UTC)

Yes, "2" is basically zero here and leading "2" doesn't mean anything like in math. Ladsgroup (talk) 06:27, 15 April 2019 (UTC)

UrlShortener support for Code Search links

Copied from -core;

14:12 (xSavitar) I'm not sure but do we think UrlShortener extension could come in handy for codesearch links?
14:13 (xSavitar) Usually, CS links are long and putting them in commit messages is not so cool per wrapping
14:13 (xSavitar) Maybe the shortener can support codesearch links? Maybe Ladsgroup can throw a word or two

This could really come in handy as CS links are usually long. Cc @Ladsgroup X-Savitar (talk) 13:16, 18 April 2019 (UTC)

That should be addressed as part of m:Talk:Wikimedia_URL_Shortener#Toolserver It's up to core platform and security team to approve it and then it's just a config change. Ladsgroup (talk) 11:11, 19 April 2019 (UTC)

Okay, thanks! X-Savitar (talk) 17:27, 19 April 2019 (UTC)

Help page

download link broken

I get the following error when trying to download this extension:

No such extension "UrlShortener".

Unable to fetch extension list!

Nicole Sharp (talk) 17:51, 6 June 2019 (UTC)

It seems to work now. (Using the default version selected). Perhaps it was just a temporary problem? Quiddity (WMF) (talk) 16:45, 7 June 2019 (UTC)

Error

This is what appears:

[c99ad0719e66fffd1396db68] Caught exception of type Wikimedia\Rdbms\DBQueryError 41.254.66.119 (talk) 18:27, 3 April 2021 (UTC)

Please see How to debug for obtaining the full error message, what you provided really isn't enough to go on. Legoktm (talk) 07:50, 4 April 2021 (UTC)

API doesn't seem to test for the URL's domain....

I've set up a script that makes use of the API to shorten URLs right from a spreadsheet, and I was surprised to find that the $wgUrlShortenerAllowedDomains is not tested then. I could shorten just about anything as long as it was a valid URL. BertrandGorge (talk) 12:42, 9 December 2021 (UTC)

Just added a bug https://phabricator.wikimedia.org/T297382 BertrandGorge (talk) 12:45, 9 December 2021 (UTC)

Request new feature for extension

Hello, everyone

If user set custom URL in short domain it looks like UrlShortener requires new parameters

Third request feature (Set Forbidden URL characters)

But add setting parameters for forbidden URL characters address by regex rule. Beginneruser (talk) 23:02, 29 December 2021 (UTC)

Now redirects to Special:UrlShortener

My site uses https but UrlShortener is saving http URLs to usc_url

How can I change this? Thanks. Jonathan3 (talk) 22:00, 24 August 2023 (UTC)

This is on purpose, it saves one character in the database. UrlShortener should be transforming it back to HTTPS on output, is that not the case? Legoktm (talk) 08:00, 25 August 2023 (UTC)

Yes that seems to work fine! (I only noticed the http when checking a short URL that refuses to redirect properly for some reason.)

Not a UrlShortener question as such, but is there something I am missing in LocalSettings.php that would obviate the need for the second 301 below?

Status code	Scheme	URL
301	https	https://www.short.uk/xyz
301	http	http://www.long.co.uk/index.php?title=Special:UrlRedirector/xyz
301	https	https://www.long.co.uk/index.php?title=Special:UrlRedirector/xyz
200	https	https://www.long.co.uk/Pagename

Jonathan3 (talk) 13:01, 25 August 2023 (UTC)

I've got one short URL that refuses to redirect properly

“Bad” IP

How necessary is it to totally restrict readers with suspicious IPs? This feature is for readers, any reader can get so IP. The IP problem will only get worse. Perhaps it will be enough to request a captcha (only for those coming from the sidebar with such IP) or remove this restriction. See also meta:Talk:No open proxies/Unfair blocking, [1][2][3], en:Wikipedia:Blocking IP addresses#Problems and solutions. @OVasileva (WMF), @Ladsgroup, @Krinkle. @Pppery, started a new topic here, as you advised Proeksad (talk) 09:18, 6 November 2023 (UTC)

While I understand, the short urls have potential to become abuse vectors, e.g. to be used to disclose someone's identity. That's why it's not allowed to create a short url while blocked. The focus should be to reduce unfair block of IPs tbh. Ladsgroup (talk) 09:57, 6 November 2023 (UTC)

I’m not sure that this helps in any way specifically for Wikimedia sites, where statistics and data on those who used the link are not provided for users.

I’m not sure that the fight against unfair blocking will lead to anything serious: this problem may not be solved for years, IPs are now being resold very actively. Moreover, since IP masking is planned, the blocking will only get stronger, in my opinion. A separate problem, as far as I understand, is that many people use proxies regularly. Proeksad (talk) 10:24, 6 November 2023 (UTC)

In the abstract, preventing blocked users from performing on-wiki actions is a reasonable anti-abuse measure. I don't think you'd disagree with that?

Anyways, agreed with Ladsgroup here. The problem is that "blocked users" includes a number of people who (presumably) have done nothing wrong and are just being blocked for their IP address, VPN, etc. That is the central issue that needs to be solved and trying to address it feature-by-feature is going to be a mess and endless rehashing of the same discussion. Adding workarounds just reduces the incentive and pressure to solve it properly. Legoktm (talk) 18:01, 6 November 2023 (UTC)

The key difference is that it is a tool for readers.

Readers are generally not restricted no matter what IP they have or whether they use a VPN. Besides, they won't even understand what it is Proeksad (talk) 07:02, 8 November 2023 (UTC)

Special:QrCode doesn't exist