Poisoned Google Image Searches Becoming a Problem
Orome1 writes "If you are a regular user of Google's image search, you might have noticed that poisoned search results have practically become a common occurrence. Google has, of course, noticed this and does its best to mark the offending links as such, but they still have trouble when it comes to cleaning up its image search results."
I was looking up images for a VP shunt when I came across a few poisoned links. I got scared for a minute because just hovering over the image triggered the payload for one of them
Two weeks ago I put some screenshots of what it looks like on my blog:
http://cobbaut.blogspot.com/
European Linux user, living in Antwerp
As a professional web developer, we often write code that expects Javascript to work on our sites, because noone ever turns it off. We have some very high traffic sites, and outside of web crawlers, I don't believe we've seen it blocked, ever.
Ironic, given that Google recently (this month) just changed its behavior to practically require Javashit.
Old hotness: (1) Google "foo". (2) Click "Images" tab at top of screen for a GIS for "foo".
New and busted: (1) Google "foo". (2) Click "Images" tab at top of screen for... "Your search - foo - did not match any documents." (3) curse, click "Images" tab again - to go to http://www.google.com/imghp?hl=en&tab=ii, and (4) have to type "foo" again in order to GIS "foo". (Or remember to start at images.google.com, which is an issue when you might not be sure which terms to use when searching for the image in the first place)
Turn Javashit on, and clicking the tab works just fine... but whatever Google changed broke the non-Javashit version of GIS.
Sorta like last month - maps.google.com is an AJAX app, so it's reasonable for it to require Javascript. But it used to work fine without cookies enabled. Now, it requires both Javascript and cookies. Interesting.
Just tested/confirmed both of these on Firefox 3.6.16.
What Facebook does overtly, Google does by benign neglect and failure to regression-test. What's next? Google services simply stop working for Firefox and require Chrome?
My wife got bitten by this just today.
She navigated to a web page from a Google search result, and Safari automatically downloaded some malware and executed it.
I didn't believe my wife's story at first, so I tried it. Sure enough, automatic download and execution on Mac/Safari.
What the fuck, Apple and Safari?
The only question that remains is whether I'll be moving her to Firefox or Chrome...
Firefox + FlashBlock + NoScript
What's the point? NoScript is FlashBlock and then some.
The trouble is that you likely get a substantially degraded experience on some sites. Many well developed sites use AJAX to speed up navigation[1], falling back on a full request when JavaScript is disabled. Similarly many sites implement convince features like jquery-based auto-completion which help make the site easier/faster to use, but again the site continues to function even with JavaScript turned off. You likely never even realize that you are getting a degraded experience because the site did not completely break.
That is a large part of the reason I actively do not recommend NoScript or similar solutions, favoring blacklisting known bothersome scripts, and using sadboxes and equivalent to guard against the unknown.
[1] You only need to download the changed portion, and browsers can update a page in place faster than re-rendering the whole page.
Stylish sheet to fix many problems in Slashdot's D3: https://gist.github.com/801524