This appears to be caused by a problem with onclick handlers. In this specific case, 
the JavaScript evaluates to false, but the link is launched anyway.

If you run the JavaScript manually, it works fine.

The specific JavaScript statement is something like:

Tabset.getInstance("friend_pills").selectTab("friend_pills_recent",function() { 
Friends.refreshView('recent'); return false;; }.bind(this),null);

If you run that in the JavaScript console it will work and execute properly.

I'll have to check to see if this fails in the general case or is specific to however 
Facebook is creating that page since that page appears to be entirely created through 
JavaScript.

Assigning all facebook.com bugs to eroman who is looking at it.

The bug is in how v8 implements the "in" operator, for negative array indexes:

  const p = -10;
  var myArray = []
  myArray[p] = [];

  // FAIL: not work in v8
  alert((p in myArray)); // Should alert true
  alert(!!myArray[p]); // Should alert true.

Facebook rewrites the onclick handlers on the page to give first chance to their own 
global event handler. They have written their own event dispatch layer which uses 
priority sorted queues of event handlers -- some of the queue priorities are 
negative, and they test for existence of a bucket using the "in" operator!

So this can cause the wrong priority version to get fired, and then nothing useful 
happens.

Note that this affects lots of links on facebook.com, not just the friends display.

Ager is working on a v8 fix.

Fixed in v8 bleeding_edge revision 117.

The bug affects also "Join this group" button, from what I can see.

it affects a whole lot of stuff on facebook, most of it casual users won't even 
notice because facebook was made to be usable without JS... no need to write every 
single thing, especially since the bug has been fixed (like comment #5 says)... wait 
til the bugfix gets packaged into a new Google Chrome build and we'll see if its 
working or not

But how do I download 

v8 bleeding_edge revision 117

?

Mr.Dooman, you'll have to wait until the next update.

does anyone have any info about expected release cycles? 

> But how do I download 
> v8 bleeding_edge revision 117

Generally what you would do, is download one of our TOT builds from http://build.chromium.org

Search the page for "[download]" and there is a link to download a .zip.
(To run side by side with existing version of chromium it is best to create a 
shortcut to the downloaded exe and add a flag like: --user-data-dir=c:\tmp\test-
profile)

That said, the bleeding_edge branch changes haven't been merged in yet, so the chrome 
builds won't see this change.

You can preview the change by checking out the source-code and patching the fix.
One way to do this is to edit the
src/DEPS file and change these lines:

  "src/v8":
    "http://v8.googlecode.com/svn/trunk@101",

to:

  "src/v8":
    "http://v8.googlecode.com/svn/branches/bleeding_edge@117",


> does anyone have any info about expected release cycles? 

I don't know the answer to this, but philosophically our goal is to do updates 
frequently and often, to fix these kinds of bugs.

Being the first update, I don't have a good sense of how long it will take.
I hope it is "soon" (<2 weeks), since this is a common problem.

Verified in build 0.2.149.28, this bug has been fixed.

Are updates going to be forced on the user? I wish a browser would force their users 
to upgrade automatically, its a lot easier to develop websites when you know everyone 
is using the same version of a browser.  When web sites get updates, users to have 
the ability to keep using an older version, I wish applications were more like web 
sites where it isn't an option to use the old code, its just upgraded automatically.

google could potentially do that because the application is installed into the 
AppData folder by default (instead of Program Files) and I don't remember the 
installer asking for elevated privileges... i would also like to see this happen, as 
transparently as possible.. the only reason I ever say no to Firefox updates is 
because I need to get work done *right now* and not go chase down compatible 
extensions.. if this can be done behind the scenes, I'm all for it

on the other hand, I'm sure you'll get a whole lot of backlash over it from the 
general public... they're already accusing Chrome of being a "keylogger" (rotfl)

This particular issue is now resolved, along with several other similar issues with 
Facebook. Thanks to the team for a very rapid resolution.

Using Help|About runs an automatic check for the latest version, and downloads the 
latest version if required. Forcing updates on users seems a bit draconian to me, 
and I think the way Chrome handles this works well.

The friends for sale app on facebook still has issues

(No comment was entered for this change.)

Closing this meta bug, as the fix for one of the major problems has been pushed out 
in the update (29)

Please file separate bugs for additional problems.