Last modified: 2013-04-08 22:46:40 UTC
At the moment, '[http://www.example.com]' is recognised as an external link by the wikimedia software, while '[HTTP://www.example.com]' is not, while my browser recognises both (even '[HtTp://www.example.com]'). I think that the protocol-recognition should be caseINsensitive. Note: I noticed this, since the linkwatcher (LiWa3, see e.g freenode #wikipedia-en-spam) that I am operating does recognise uppercase HTTP as a protocol, and found an edit (http://en.wikipedia.org/?diff=417579535) where a blacklisted url (see http://meta.wikimedia.org/wiki/Spam-blacklist) was added to a page (but it is not a 'clickable link'). In fact, changing the capitalisation to lowercase disallows the edit. Therefore, this should be propagated into the blacklist code as well, if necessary.
RFC 3986 [1] states that schemes are case-insensitive but canonical form is lowercase. The parser should normalize the scheme part to use lower case. Depending on Spam Blacklist hook point, it might or might not need an adaptation to take care of upper case schemes. [1] http://tools.ietf.org/html/rfc3986#section-3.1
(In reply to comment #0) > At the moment, '[http://www.example.com]' is recognised as an external link > by the wikimedia software, while '[HTTP://www.example.com]' is not Cannot reproduce on test2 with 1.22wmf1 - both seem to render as external links (likely also because [ is used insteadof [[ ?). How can I see this problem?
Hmm, it seems that it has changed somewhere automagically then. This was a problem 2 years ago ... By the way, the single brackets are necessary for an external link, double brackets are for wikilinks. I am changing the status of the bug .. seems 'resolved'
bug 34939 was a duplicate of this bug and got fixed with https://gerrit.wikimedia.org/r/#/c/15224/
