Post a Comment
"aka. normal proper testing"
In an ideal world yes, in the real world no: nearly all the unit test I've seen check the behaviour of the application when given "normal" data.
Given this, it makes sense to use another name for 'security testing' where you're explicity testing the application against a cracker.
Of course 'security testing' is not restricted to fuzzing..
I've found that to be the case even when testing web-based forms. Ten percent of the time is spent making sure the forms work when users fill them out correctly; the other ninety percent of the time is occupied by making sure there is proper error-handling when users enter information incorrectly.
One such tool are zzuf, http://sam.zoy.org/zzuf/
For testing browsers this is not anything new, here is one article about it from back in April 2006. http://www.theregister.co.uk/2006/04/13/data_fuzzing/
You can run the test on your browser of choice, to see how long it takes before it crash :-) http://metasploit.com/users/hdm/tools/see-ess-ess-die/cssdie.html
In case you were actually being serious, no this will not help find memory leaks.
I'm pretty sure that was supposed to be sarcasm.
They're working on them, and the goal is for the next version to be completely free. One of the developers recently took Firefox3 to the 500 most popular internet sites, and there was only 1 leak (which was immediately fixed). Of course, he didn't do any navigation inside the sites, only going to the main page. But that is still pretty good.








