[MC-8437] URL parser does not detect 4 letter TLDs

The URL parsing in chat dislikes Top Level Domains with more than 3 characters (eg a .com works fine, but a .info does not), and does not permit clicking on the link in game to take the user to a web address.

The problem is easily reproduced, after trying the domains in the chat box:

jameo.com and other variations (eg prepending http://) functions correctly.

jameo.info and other variations do not.

Linked issues

relates to 2

MC-15077 Clickable links in Chat are broken for many valid URL formats Resolved

MC-18898 URL parser will not detect TLDs of over 4 characters Resolved

Comments 4

bugi74 2013-01-26T16:18:36Z

Bug confirmed, found and busted. Erm, fixed.

Bug

"ChatClickData"

public static final Pattern pattern = Pattern.compile("^(?:(https?)://)?([-\\w_\\.]{2,}\\.[a-z]{2,3})(/\\S*)?$");

Fix

"ChatClickData"

public static final Pattern pattern = Pattern.compile("^(?:(https?)://)?([-\\w_\\.]{2,}\\.[a-z]{2,4})(/\\S*)?$");

Fix tested on 1.4.7

(Edit: minor technical detail: It does allow clicking the text in any case, it just won't recognize it as an URL without that small change, and thus will not react any further.)

kumasasa 2013-01-26T23:10:12Z

Confirmed.

MiiNiPaa 2013-01-27T15:11:48Z

http://en.wikipedia.org/wiki/List_of_Internet_top-level_domains
Not every TLD will catched by this RegEx. You should avoid "TLD can be only 2-4 characters long" type of thinking.

kumasasa 2013-01-27T15:17:45Z

... there are as many regex as there are URLs : http://regexlib.com/%28X%281%29A%28w_EUcf1Bu6N3WFkuVKjGepFvJE2idKTn6XdSnjWUuNtd1un_Eb4FoDkOVAR8SRiMGGBF5Tbw1UonIHucQzOcdTgoYTvkxhpmUooMy90R-WvvA2aGvfMrruwlVe85XxuakVN09q6CjmodpNV4bx0z13Hec3uqWgCLn6W8bpOnukXFFUETJvMR0rkxKy7Z6EVW0%29%29/Search.aspx?k=URL&c=-1&m=-1&ps=100