[whatwg] select element should have a required attribute

[Garrett Smith]

On 6/19/10, Garrett Smith <dhtmlkitchen at gmail.com> wrote:
> On 6/19/10, Ashley Sheridan <ash at ashleysheridan.co.uk> wrote:
>> On Fri, 2010-06-18 at 13:34 -0700, Garrett Smith wrote:
>>
>>> On 6/18/10, Garrett Smith <dhtmlkitchen at gmail.com> wrote:
>>> > On 6/18/10, Ashley Sheridan <ash at ashleysheridan.co.uk> wrote:
>>> >> On Fri, 2010-06-18 at 11:35 +0200, Mounir Lamouri wrote:
>>> >>
>>>
>>> [...]
>>> >
>>> > Implementations vary on actual behavior. See also:
>>> > https://bug-20184-attachments.webkit.org/attachment.cgi?done=true&id=22519
>>> >
>
> That is the wrong URL to begin the test; instead, the following URL
> must be used:
> https://bug-20184-attachments.webkit.org/attachment.cgi?id=22519
>
> Looking closer, that test is no good. Doesn't run in IE anyway. Since
> I made it, it's my fault and so I've attached a simpler test bug. The
> simpler test is still not that good because it tests two things:
>   1) is a SELECT with no name successful?
>   2) is a disabled, selected OPTION in a SELECT successful?

[...]

> Controls that are disabled cannot be successful. Browser behavior
> varies on what happens when a SELECT has a disabled OPTION selected.
>

Although the test uses PASS and FAIL, I'm not sure that's right.

What is the correct expectation for the outcome of submitting a SELECT
that has a disabled, selected OPTION?

The SELECT itself should be successful, but since the selected OPTION
is disabled, does that make the SELECT unsuccessful? Looking at the
spec, it seems that that shouldn't change anything. You may be right,
and perhaps the test is still wrong.

Recent browsers vary.

Does HTML 5 cover this case?

Garrett