Making sure http appears in a web address

I have people posting their website address but variations are posted such as:

• theirsite.com
• www.theirsite.com
• http://theirsite.com
• http://www.theirsite.com

When I link to an address without http:// it takes the link as internal

<a href="theirsite.com">their site</a>

sending people to something like: http://mysite.com/thiersite.com

Another option I've tried is linking to something like mysite.com/?link=theirsite.com - This way I can do some link tracking etc then redirect people to the link but it has the same problem:

//do some tracking etc here
$link =$_GET['link'];
header("Location: $link");

From stackoverflow

• I would provide some validation or sanitation. Use a regex to see if http:// starts with it. If it doesn't, either throw an validation error or put http:// at the start.

• if not "://" in users_url:
      users_url = "http://" + users_url
  

  ...or equivalent, in the language of your choice.

• You could use regular expressions to test input

  Regex exp = new Regex(
      @"http://(www\.)?([^\.]+)\.com",
      RegexOptions.IgnoreCase);
  

  OregonGhost : Hopefully the .com is just an example :D

• put "http://" in the field by default, then validate the URL with something like

  if(eregi("^((http|https)://)?([[:alnum:]-])+(\.)([[:alnum:]]){2,4}([[:alnum:]/+=%&_.~?-]*)$", stripslashes(trim($_POST['link'])))){
      //link is valid
  }
  

  if link does not validate, just print them a message saying "the link you entered is invalid, make sure it starts with 'http://'"

  Fredrik Mörk : I like how this will also take care of the cases when people write "no" or "not yet" or stuff like that.

  Gumbo : +1 For the default value for the URL field. But I wouldn’t use `eregi`.

  Peter : good point with people posting "no" or "not yet"...

  The Pixel Developer : No need to use regular expressions in this case. PHP has built in URL validation using the "filter" extension. See my answer below.

  vartec : ereg is dead, you should use PCRE instead.

• I would use something like this:

  $link = str_replace(array("\r", "\n"), '', trim($link));
  if (!preg_match('/^https?:\/\//', $link)) {
      $link = 'http://'.$link;
  }
  header('Location: '.$link);
  

  Another way would be the parse_url function to parse the given URL, see what parts are missing and add them.

  Peter : might need to consider trailing space as noted by Elazar and also there was a small typo repalce. Cheers.

  Gumbo : Have you noticed the `trim`?

  Peter : oops, silly me!

• Please note, there's a real difference between www.site.com and site.com, usually both works, but on some website each leads to a different path (some badly defined website won't work without the www for instance). So You can't always prepend 'www' to the input.

  Another note, do handle prepending space, so that ' http://' would not be prepended with additional http://.

  My Javascript Regex based solution

  'http://'+field.replace(/^ *http:\/\//,'')
  

  You can verify that on the client size, just put a code in similar spirit on the onSubmit of your form.

• No need to use regular expressions here. PHP has URL validation built in.

  Filter Var

  var_dump((bool) filter_var('http://www.website.com', FILTER_VALIDATE_URL, FILTER_FLAG_HOST_REQUIRED));
  var_dump((bool) filter_var('http://website.com', FILTER_VALIDATE_URL, FILTER_FLAG_HOST_REQUIRED));
  var_dump((bool) filter_var('www.website.com', FILTER_VALIDATE_URL, FILTER_FLAG_HOST_REQUIRED));
  var_dump((bool) filter_var('website.com', FILTER_VALIDATE_URL, FILTER_FLAG_HOST_REQUIRED));
  

  Output

  bool(true)
  bool(true)
  bool(false)
  bool(false)
  

  Please do not jump straight to regular expressions for validation, PHP has a lot of methods built in to deal with these scenarios.

  -Mathew

  Gumbo : But regular expressions are language independent.

  The Pixel Developer : To be fair, he posted it with a PHP tag and I've given him a PHP answer. It's a non-issue.