W3C home > Mailing lists > Public > public-html@w3.org > February 2008

Re: [WF2] HTMLFormElement.accept vs. function accept(){}

From: Jon Barnett <jonbarnett@gmail.com>
Date: Wed, 20 Feb 2008 16:02:56 -0600
Message-ID: <bde87dd20802201402g681e73f6yd32cc2376fa16034@mail.gmail.com>
To: "Simon Pieters" <simonp@opera.com>
Cc: liorean <liorean@gmail.com>, public-html <public-html@w3.org>

On Wed, Feb 20, 2008 at 12:36 PM, Simon Pieters <simonp@opera.com> wrote:
>
>
>  On Wed, 20 Feb 2008 19:06:33 +0100, liorean <liorean@gmail.com> wrote:
>
>  >
>  > On 20/02/2008, Simon Pieters <simonp@opera.com> wrote:
>  >> There are some ways to fix this problem.
>  >>
>  >>    1) Make the .accept DOM attribute not override functions in scope.
>  >>    2) Remove support for .accept DOM attribute but keep the content
>  >> attribute.
>  >>    3) Remove support for both .accept and <form accept>.
>  >
>  > 4) Give it a different name in the DOM bindings, such as
>  > formelm.htmlAccept.
>  >
>  > This is style with other bindings used to avoid script engine
>  > problems, such as legendelm.htmlFor or elm.className. Granted, both
>  > those are because of ECMAScript reserved keywords and not live content
>  > script usage collisions, but I still think it's a viable option.
>
>  Yeah... if <form accept> was a killer feature. But it's not. It's pretty
>  useless. DOM2 HTML forgot about it's existence altogether. AFAIK, no
>  browser does anything with it at all. Supporting it is causing Web compat
>  problems. I'd say it's ripe for removal.
>
>  --
>  Simon Pieters
>  Opera Software
>
>

As an author is easy enough to fix by changing the accept() function
call to window.accept() since globally defined functions end up as
methods of the window object.

As an author, that type of hell is not new.  A form field named "name"
or even "action" is hell to deal with, especially in IE where
form.getAttribute() actually returns the input element instead of the
attribute value.  Try this in IE6:
http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Cform%20action%3D%22foo%22%20onsubmit%3D'alert(this.getAttribute(%22action%22))%3B%20return%20false%3B'%3E%3Cinput%20type%3Dsubmit%3E%3Cscript%3E%20document.forms%5B0%5D.firstChild.click()%20%3C%2Fscript%3E

It's a problem that's unavoidable with any new properties/methods on
HTMLFormElement.  Since it's easy to work around and common to work
around, is it worth fixing?

-- 
Jon Barnett
Received on Wednesday, 20 February 2008 22:03:12 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:16:12 GMT