Skip to main content

Browse the archive

Revision 156a37a385f8c27d8b35995e88b913d5cfb1e02a authored by Simon Sapin on 21 April 2016, 08:35:40 UTC, committed by Anne van Kesteren on 21 April 2016, 08:35:40 UTC 
Update websocket test: C0 controls are trimmed from URLs
 
https://html.spec.whatwg.org/multipage/comms.html#dom-websocket says to parse the first argument with https://url.spec.whatwg.org/#concept-url-parser, which in turns says:

> Remove any leading and trailing C0 controls and space from input.

Which links to:

> The C0 controls and space are C0 controls and code point U+0020. 
> The C0 controls are code points in the range U+0000 to U+001F, inclusive.

So a trailing U+0000 is removed from a string when parsing it as an URL.
1 parent 3d40f9b
Raw File
open-url-bogus.htm 
<!DOCTYPE html>
<html>
  <head>
    <title>XMLHttpRequest: open() - bogus URLs</title>
    <script src="/resources/testharness.js"></script>
    <script src="/resources/testharnessreport.js"></script>
    <link rel="help" href="https://xhr.spec.whatwg.org/#the-open()-method" data-tested-assertations="following::ol/li[7] following::ol/li[8]" />
  </head>
  <body>
    <div id="log"></div>
    <script>
      function url(url) {
        test(function() {
          var client = new XMLHttpRequest()
          assert_throws("SyntaxError", function() { client.open("GET", url) })
        }, document.title + " (" + url + ")")
      }
      url("http:")
      url("http://a a/")
    </script>
  </body>
</html>
back to top