On 3/3/12 5:06 PM, R wrote: > Thanks Conrad, much cleaner indeed. But, this does not solve my > invalid internet URL issues.... still accepts pretty much any URL.
I don't understand what you mean by "invalid." I would expect that NSDataDetector has been pretty thoroughly tested and will not match syntactically invalid results. If you find otherwise, you should file a bug report. I suspect that the disconnect here is that *your* definition of invalid is closer to "looks funny" than to "violates RFC 1738." I certainly hope that Apple is going for the latter. Let's consider the example from your original post: http://www.cnn.thePiratesWillWinTheWorldSeries I see absolutely nothing wrong with that URL. Sure, it doesn't have a TLD at the end. But that has nothing to do with being valid. If you don't believe me (and don't want to mess around with your DNS server) add the following line to /etc/hosts: 157.166.255.18 www.cnn.thePiratesWillWinTheWorldSeries Safari will now bring up CNN if you punch in your URL. Here are a few other more common examples of URLs that would break if you try to invent your own validation scheme in contravention of standards: http://localhost http://127.0.0.1 http://[2001:470:1f0e:a20::2] (You *are* supporting IPv6, right?) Now, this situation is going to be made even more complicated by the proliferation of new gTLDs (see, e.g., http://en.wikipedia.org/wiki/Generic_top-level_domain). You can bet we are soon going to see www.iphone.apple as a usable, if perhaps not advertised, URL! Oh, and http:// is far from the only acceptable protocol. (https://? ftp://?) Hopefully I have convinced you that this is a losing proposition. There is a general rule when developing network software, or perhaps all software: be liberal in what you accept, conservative in what you emit. > I'm wondering how Twitter IOS, and TWTweetComposeViewController > handles URL counting. I don't know, but I would be upset if a valid URL wasn't recognized properly. As for determining whether a URL is valid, you could do a hostname lookup (if the URL isn't an IP itself). But then you have several issues: 1) A minor point, but (as with all operations that potentially touch the network) you need to make sure to do this on a background thread. Host lookups can return in an indeterminate amount of time. You would probably want to have a timeout and then fall back to a different mechanism. 2) If you are in, say, a dual-horizon DNS environment, your hostname might resolve in one part of a network and not in another. This would give inconsistent results as a user moves around. 3) If a domain's authoritative DNS server is down (rare but far from unheard of), you might temporarily get a failure condition. This would *really* confuse and irritate the user ("sometimes the URL shortening works, sometimes it doesn't, for the same site!"). As you can see, this is far from a straightforward problem. I'd jettison this notion entirely and just accept anything that the pertinent RFC(s) tell you to. ` -- Conrad Shultz Synthetiq Solutions www.synthetiqsolutions.com _______________________________________________ Cocoa-dev mailing list (Cocoa-dev@lists.apple.com) Please do not post admin requests or moderator comments to the list. Contact the moderators at cocoa-dev-admins(at)lists.apple.com Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com
