[NavigationTiming] Uniqueness requirement for PerformanceTiming doesn't make much sense

The specification curently says (section 4.4):

   Each browsing context must have a unique window.performance.timing
   attribute.

Since browsing contexts outlive page loads, this means, on its face, 
that the PerformanceTiming object is the same across pageloads in the 
same browsing context.

I just tested, and neither Chrome nor Gecko have that behavior (tested 
via direct == check).

IE9 won't let me examine PerformanceTiming objects associated with 
unloaded pages (even as far as testing whether they're == to something, 
and even if I save the object in a variable before the page is 
unloaded), which presumably means that they're distinct objects from the 
ones of currently-loaded pages.

So none of these do what the spec calls for, which is good because what 
the spec calls for makes no sense to me.

What's this text actually trying to get at?

-Boris

P.S.  In case this is not clear, this stuff is all in the 
"implementation experience" bucket for this spec for Gecko; I'm looking 
through our implementation and comparing it to what the spec actually says.

Received on Sunday, 24 June 2012 05:41:26 UTC