Monthly Archives: August 2010

PHP OAuth extension: trial, error and success

I’ve been experimenting with the PHP OAuth PECL extension over the last few days and initially ran into some small problems getting it to function correctly. So for the sake of making this world wide web an even better place, here are some error-messages you might encounter and what you could do to resolve them:

  • “OAuth::getRequestToken() URL file-access is disabled in the server configuration” → the OAuth extension by default uses fopen (streams) to fetch data from the OAuth provider (server), but you’re probably slightly paranoid (or is it security- conscious?) and have allow_url_fopen = 0 in you php.ini. Just change that value to “1” and reload your webserver and then see …
  • SSL: fatal protocol error → OAuth seems to be working fine, but for reasons unknown OAuth insists something fatal just happened. You could ignore this one, or try to lower the error-reporting value or do as I did and decide to use OAuth::setRequestEngine to switch from those doomed streams to almighty Curl, until …
  • “setRequestEngine expects parameter to be long, string given” → so you installed (compiled) the OAuth-extension, but you didn’t have that libcurl-dev package installed, did you? No matter how polite you request that Curl-engine, it is just not going to happen unless you install libcurl-dev, and then re-installed OAuth.

And after successfully re-installing OAuth, this time with Curl-support, you could try to build a small application, accessing GMail maybe?

As found on the web (August 4th)

blog (feed #46)
generic (feed #49)
facebook (feed #40)
frank Frank vraagje voor wie al in sinescoop is geweest; is dat daar wat te doen qua akoestiek, drukte, … of trekken we toch beter naar metropolis?.
generic (feed #49)
youtube (feed #51)
frank liked Zaz – Je veux.
youtube (feed #51)
blog (feed #46)
frank published WordPress stats oddity.
generic (feed #49)
generic (feed #49)
blog (feed #46)

iGoogle Facebook gadget security flaw fixed & explained

I just received confirmation from the Google Security Team that the bug I discovered in the iGoogle Facebook Gadget which allowed attackers to log into an other user’s Facebook account bypassing all authentication, has been fixed. So now that the hole has been closed, let’s look at what was happening, shall we?

The gadget uses the Facebook’s Javascript API to the connect with Facebook, asking you for permission to access your FB data. In the process of getting that authorization, the gadget exchanges tokens with Facebook, some of which should absolutely be kept safe from prying eyes. And that’s where things went wrong: the gadget had the authentication info in the URL. So if a user of the iGoogle Facebook gadget clicked a link to an external site in the news feed, the request for that page had a referrer that contained all authentication-info.

And that’s exactly what happened on last week, when I spotted this referrer in my blog stats:, %22session_key%22:%2291d52d2ed5a130fd941b11f1-1175373488%22, %22secret%22:%22fdee68961b3cdee5b51390a4bdeac7a0%22,%22expires%22:0, %22access_token%22:%2283101558C90fd9KfA9KJQh5uT98TqIjxQpzUi4.%22,

You can guess what happened when I opened that URL; the iGoogle Facebook gadget initialized using the embedded credentials, automatically logging me in as the guy that was unlucky enough to have clicked the link to my blog.

But how could this vulnerability have been exploited, you may ask? Well, easy enough; create a page that is viral enough for people to share or like  (likespam or even likejacking) and wait for users of the iGoolge Facebook-gadget (there’s over 1 million of them after all) to follow the links, feeding your webserver logfiles with credential-rich referrers.

As Google confirmed this bug indeed has been fixed. The new version of the gadget, which was deployed late last week, does not leak credentials in the referrer-URL any more:,iHKb-4mKuMY/lib/librpc.js,vrFMICQBNJo/lib/libcore.js,a5j4V1JuNVE/lib/libsetprefs.js

So if anyone asks me what my good deed for this year was; I helped protect 1 million people’s Facebook accounts from being hacked.

Sounds swell, no? ;-)

WordPress stats oddity

A couple of months ago I removed Google Analytics from this wee little blog, but I still use the less sophisticated stats plugin to follow up on what is being read around here. Loading the stats-page in my Android-browser is far from optimal (it uses Flash to draw the charts and it is a pretty big page), so I was pleased to read that version 1.3 of WordPress for Android features a stats-section. But the reports don’t work, I just get “No stats data found, please try again later”.

Now while playing around with the stats API over the weekend, I noticed some unusual things:

  • “” works
  • “” results in “Error: zero rows returned.”

The API also supports blog_id instead of blog_uri and after some digging for blog_id’s I found that they are listed in (the html source of) the blog dropdown-list on your stats page. And there the problem became clear: I had two blog_id ‘s for the same blog_uri ( and the first one was defunct:

  • “blog_id=2184847” results in “Error: zero rows returned.”
  • “blog_id=2185033” works just fine

As I can’t remove the entry with the faulty blog_id from the Stats DB and as I can’t change the Android WordPress app to use one of multiple blog_id’s instead of the blog_uri, I can’t fix this little bugger myself, so I contacted WordPress Support. But how did I end up with 2 blog_id’s in the database?