Re: navigationStart in NavigationTiming

Glad to hear the implementations are consistent. Does anyone have a
suggestion for what the definition of navigationStart should be?

I struggled with defining it, and after reading
http://www.w3.org/TR/html5/browsers.html#browsing-context, specifically
this line there,
"When a browsing
context<http://www.w3.org/TR/html5/browsers.html#browsing-context> is
first created, it must be created with a single
Document<http://www.w3.org/TR/html5/dom.html#document> in
its session history, whose
address<http://www.w3.org/TR/html5/dom.html#the-document's-address>
 is about:blank <http://www.w3.org/TR/html5/infrastructure.html#about:blank>
"

I wonder if the case where there is no previous document even exist
theoretically. So I suggest changing the following 3 definitions in the
spec (in all of them, I just remove the clause "if there is no previous
document"):

navigationStart attribute

This attribute must return the time immediately after the user agent
finishes prompting to
unload<http://www.w3.org/TR/html5/browsers.html#prompt-to-unload-a-document>
the
previous document.
unloadEventStart attribute

If the previous document and the current document have the same
origin<http://tools.ietf.org/html/rfc6454>
 [IETF RFC 6454<http://www.w3.org/TR/2012/REC-navigation-timing-20121217/#rfc6454>],
this attribute must return the time immediately before the user agent
starts the unload<http://www.w3.org/TR/html5/browsers.html#unloading-documents>
event
of the previous document. If the previous document has a different
origin<http://tools.ietf.org/html/rfc6454> than
the current document, this attribute must return zero.
unloadEventEnd attribute

If the previous document and the current document have the same same
origin<http://tools.ietf.org/html/rfc6454>,
this attribute must return the time immediately after the user agent
finishes the unload<http://www.w3.org/TR/html5/browsers.html#unload-a-document>
event
of the previous document. If the previous document has a different
origin<http://tools.ietf.org/html/rfc6454> than
the current document or the unload is not yet completed, this attribute
must return zero.

If there are HTTP redirects or
equivalent<http://www.w3.org/TR/html5/infrastructure.html#concept-http-equivalent-codes>
when
navigating and not all the redirects or equivalent are from the same
origin<http://tools.ietf.org/html/rfc6454>,
both unloadEventStart<http://www.w3.org/TR/2012/REC-navigation-timing-20121217/#dom-performancetiming-unloadeventstart>
 and unloadEventEnd<http://www.w3.org/TR/2012/REC-navigation-timing-20121217/#dom-performancetiming-unloadeventend>
must
return the zero.


Does this make sense?

Arvind


On Wed, May 1, 2013 at 1:26 PM, Jatinder Mann <jmann@microsoft.com> wrote:

> IE10 on right clicking on CNN.com and opening it a new tab has
> navigationStart < fetchStart and the others are zero'd out as well.
>
> >> performance.timing.navigationStart
> 1367439590579
> >> performance.timing.fetchStart
> 1367439590612
> >> performance.timing.redirectStart
> 0
> >> performance.timing.redirectEnd
> 0
> >> performance.timing.unloadEventStart
> 0
> >> performance.timing.unloadEventEnd
> 0
>
> Thanks,
> Jatinder
>
> -----Original Message-----
> From: Boris Zbarsky [mailto:bzbarsky@MIT.EDU]
> Sent: Wednesday, May 1, 2013 1:01 PM
> To: Arvind Jain
> Cc: public-web-perf
> Subject: Re: navigationStart in NavigationTiming
>
> On 5/1/13 3:04 PM, Arvind Jain wrote:
> > Btw if you could also test http://cnn.com, and make sure
> > navigationStart != fetchStart, that would be great. This is an example
> > of cross origin redirect, and even here, navigationStart should be the
> > time before the redirect is initiated even though redirectStart would
> show 0.
>
> [15:59:16.137] performance.timing.navigationStart
> [15:59:16.139] 1367438338927
> [15:59:20.456] performance.timing.fetchStart [15:59:20.458] 1367438339335
> [15:59:26.576] performance.timing.redirectStart [15:59:26.578] 0
> [16:00:10.539] performance.timing.unloadEventEnd [16:00:10.541] 0
> [16:00:17.733] performance.timing.unloadEventStart
> [16:00:17.735] 0
>
> -Boris
>
>
>
>
>