Page 1 of 1

[INVALID] web bug blocking regression

Posted: Tue Nov 02, 2010 9:02 am
by al_9x
fx 3.6.12, NS 2.0.5rc1, new profile, forbid web bugs

This ad iframe has a web bug which redirects. The initial request is not blocked, but the redirect is aborted with an error:

Code: Select all

Error: [Exception... "'NoScript aborted redirection to http://ad.yieldmanager.com/imp?&t=2' when calling method: [nsIChannelEventSink::onChannelRedirect]"  nsresult: "0x8057001e (NS_ERROR_XPC_JS_THREW_STRING)"  location: "<unknown>"  data: no]
Curiously, if I save and load this iframe from localhost, the web bug is blocked and there is no redirect error.

Also, please take a look at http://forums.informaction.com/viewtopi ... =10&t=4832

Re: web bug blocking regression

Posted: Tue Nov 02, 2010 4:43 pm
by al_9x
Can you repro?

Re: web bug blocking regression

Posted: Fri Nov 05, 2010 11:28 pm
by Giorgio Maone
Not a bug.
Web bug blocking checks cross-site image loads only, for several reasons (if you think about it, same-site tracking has several more effective options than web bugs).
Initial image is loaded same-site (ad.harrenmedianetwork.com), then redirected to ad.yieldmanager.com, when it gets correctly blocked.

Re: web bug blocking regression

Posted: Sat Nov 06, 2010 3:13 am
by al_9x
Giorgio Maone wrote:Not a bug.
Web bug blocking checks cross-site image loads only, for several reasons (if you think about it, same-site tracking has several more effective options than web bugs).
Yes, that makes sense. I lost sight of the fact that it's the same host as the already loaded iframe.
Giorgio Maone wrote:Initial image is loaded same-site (ad.harrenmedianetwork.com), then redirected to ad.yieldmanager.com, when it gets correctly blocked.
Is throwing this exception the only way to block it? It's the error in the console that caught my attention, since usually that indicates a bug.

Re: web bug blocking regression

Posted: Sat Nov 06, 2010 9:05 am
by Giorgio Maone
al_9x wrote:Is throwing this exception the only way to block it? It's the error in the console that caught my attention, since usually that indicates a bug.
Yes it is. That's the contract of the redirection callback API: the redirection is aborted if an exception is thrown by its callbacks. The console noise is an unfortunate but unavoidable side-effect, since this exception obviously cannot be catch and silenced by JavaScript code.