How can I test if a string is URL encoded?
Which of the following approaches is better?
There's no reliable way to do this, as there are strings which stay the same through the encoding process, i.e. is "abc" encoded or not? There's no clear answer. Also, as you've encountered, some characters have multiple encodings... But...
Your decode-check-encode-check scheme fails due to the fact that some characters may be encoded in more than one way. However, a slight modification to your function should be fairly reliable, just check if the decode modifies the string, if it does, it was encoded.
It won't be fool proof of course, as "10+20=30" will return true (+ gets converted to space), but we're actually just doing arithmetic. I suppose this is what you're scheme is attempting to counter, I'm sorry to say that I don't think there's a perfect solution.
HTH.
Edit:
As I entioned in my own comment (just reiterating here for clarity), a good compromise would probably be to check for invalid characters in your url (e.g. space), and if there are some it's not encoded. If there are none, try to decode and see if the string changes. This still won't handle the arithmetic above (which is impossible), but it'll hopefully be sufficient.