- From: Ian Hickson <ian@hixie.ch>
- Date: Mon, 16 Apr 2007 03:35:38 +0000 (UTC)
- To: Dave Raggett <dsr@w3.org>
- Cc: "public-html@w3.org" <public-html@w3.org>
On Mon, 16 Apr 2007, Dave Raggett wrote: > > > > See: http://www.whatwg.org/specs/web-apps/current-work/#parsing > > Thanks for the pointer. As far as I could tell, you are indeed > suggesting tracking which form the fields belong to without the need for > the form element to contain them. What wasn't clear was how the form > element's start and end tags are manipulated. Search for "form" (_with_ the quotes) in the parsing section -- it's fully defined. To see what DOM you would get, try samples with the html5lib parser [1]; note that, as you point out, the form fields do not necessarily get associated with form elements in their ancestor chain. For what it's worth, this strategy is based on implementation feedback from Mozilla and Safari and is thought to be compatible with the majority of Web content with illformed <form> elements. [1] http://code.google.com/p/html5lib/ > BTW that document is *very* unwieldly and slows my browser to the speed > of a snail. I recommend breaking large documents into smaller ones and > taking advantage of the power of hypertext links. As noted at the very top of the document, there already exists a multipage version: http://www.whatwg.org/specs/web-apps/current-work/multipage/ -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.'
Received on Monday, 16 April 2007 03:35:59 UTC