1 Nov 2004 19:08
Re: New URL spoofing bug in Microsoft Internet Explorer
Russ <Russ.Cooper <at> RC.ON.CA>
2004-11-01 18:08:05 GMT
2004-11-01 18:08:05 GMT
Well, code that is improperly formed may well cause user agents to try and figure out for themselves what they should do. Code that isn't permitted, however, should IMO be handled differently. For example, an A Element implicitly denies the inclusion of a Button or Form Element, and a Button Element explicitly denies the inclusion of an A Element. The following examples all work, regardless where you put your mouse. They all show microsoft.com in the status bar, but when the text is click they all go to google.com. I contend they shouldn't. The fact that the A Element is being allowed to function is what makes any/all of these potentially harmful. If stricter interpretation rules applied to the A Element, a great deal of phishing might be avoided. Ken Grohs feels this is all a moot point, he points out that you can do this by using onmouseover and onmouseout. I agree, but he has to use script. IMO, A Elements should be perfectly formed or not work as links, display in the status bar, or render as links do (e.g. visited link color, etc...) Credit to http-equiv for making the button disappear. <form action="http://www.google.com" method="get"> <a href="http://www.microsoft.com/"> <button title="http://www.microsoft.com" type=submit style="BORDER: 0pt; CURSOR: hand; COLOR: blue; BACKGROUND-COLOR: transparent; "> http://www.microsoft.com </button> </a> </form> <form action="http://www.google.com" method="get"> <a href="http://www.microsoft.com/"> <button title="http://www.microsoft.com" type=submit style="BORDER: 0pt; CURSOR: hand; COLOR:(Continue reading)
RSS Feed