The XHR spec supports all four, w/o the _method.  I'll have to find the article 
I read a while back about it :\

On Jun 8, 2011, at 11:54 AM, gonchuki wrote:

> The issue is not if the HTTP spec supports the methods or not, it is
> the browsers that do not support those methods, and that's why the
> _method shim was created.
> 
> I tried looking for relevant references on the web and could only find
> a Prototype bug from 2008 --
> https://prototype.lighthouseapp.com/projects/8886/tickets/289-allow-put-delete-and-other-http-methods-on-ajax-requests
> where some of the comments clearly explain that there's no cross-
> browser support for verbs other than GET and POST.
> This WebOS forums post also shows at which point WebKit introduced
> support for the PUT method --
> https://developer.palm.com/distribution/viewtopic.php?f=15&t=2373&p=18572&hilit=put+request#p18572
> 
> TL;DR -- only Firefox supports PUT and DELETE, and post Feb 2010
> builds of WebKit suppoort PUT requests (not sure about DELETE).
> 
> 
> 
> On Jun 7, 2:49 pm, Ryan Florence <[email protected]> wrote:
>> Right, the xmlHttpRequest spec and HTTP spec support all four methods, while 
>> the HTML5 form attribute "method" does not.
>> 
>> On Jun 7, 2011, at 3:45 PM, Sean McArthur wrote:
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>>> HTML does not determine HTTP. Those other HTTP actions exist, regardless of 
>>> whether they're legal attributes in forms in HTML5.
>> 
>>> On Tue, Jun 7, 2011 at 1:08 PM, gonchuki <[email protected]> wrote:
>>> Actually, only GET and POST are official methods in the HTML5 spec
>>> [1]. PUT and DELETE were there at some point, but have since been
>>> removed so you shouldn't assume that any/all browsers will support it.
>>> Also, I know this holds true for most XHR "1" implementations, not
>>> sure if all the new XHR2 capable browsers implement those other
>>> methods regardless of what the HTML5 spec says.
>> 
>>> [1]http://www.w3.org/TR/html5/association-of-controls-and-forms.html#att...
>> 
>>> On Jun 6, 7:08 am, Lee <[email protected]> wrote:
>>>> On 06/06/2011 16:06, Tim Wienk wrote:
>> 
>>>>> If I understood him correctly what he wants is to actually *not*
>>>>> emulate, and use the actual "delete" and "put" methods, and the
>>>>> problem was that it emulates by default and changes your "delete" into
>>>>> a "post".
>> 
>>>> Exactly. It seems a strange default, to mange HTTP, but I can see the
>>>> potential reasoning sine you pointed out the Rails approach.
>> 
>>>> Thanks again
>>>> Lee

Reply via email to