- From: Ian Hickson <ian@hixie.ch>
- Date: Sat, 17 Oct 2009 09:12:48 +0000 (UTC)
On Tue, 13 Oct 2009, Boris Zbarsky wrote: > On 10/13/09 10:58 PM, Ian Hickson wrote: > > Ah, I see. Yes, that makes sense. Are you ok with what the spec says > > despite this minor difference? (It seems that what the spec says is a > > mote more self-consistent, but I could be biased! ;-) ) > > Can you point me to the relevant spec text? Sure. The 'beforeunload' and 'unload' events are described in algorithsm here: http://www.whatwg.org/specs/web-apps/current-work/multipage/#unloading-documents If you click the "prompt to unload the document" and "unload a document" bits you'll get links to the algorithsm that use them. In particular, the "navigate" algorithm: http://www.whatwg.org/specs/web-apps/current-work/complete.html#navigate Step 3 is the one that prevents a browsing context from being navigated while 'unload' is running. Step 8 is the one that runs 'beforeunload'. Step 17 references algorithms that each later reference the "update the session history with the new page" algorithm: http://www.whatwg.org/specs/web-apps/current-work/complete.html#update-the-session-history-with-the-new-page Step 1 of that algorithm invokes 'unload'. Sorry for linking to complete.html; the multipage copy is currently undergoing maintenance as I write this. If you replace "complete.html" with "multipage/" then you will get redirected to the multipage version. HTH, -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.'
Received on Saturday, 17 October 2009 02:12:48 UTC