Re: XMLHttpRequest send() Method Uploads On Inappropriate HTTP Protocols

On Tue, 23 Feb 2010 12:29:59 +0100, Hans Meiser <brille1@hotmail.com>  
wrote:
>>> I believe this spec ought to be updated to also exclude POST data from
>>> being uploaded for HTTP request methods like DELETE, TRACE, OPTIONS.
>>
>> As I said in the bug, we do not want to constrain HTTP more than we
>> already do.
>
> I believe this is not a question of wanting but a question of definition  
> and browser compatibility.

Indeed. And as such we are doing the right thing here.


>>> Plus, I believe upload should be allowed to be send along with HEAD
>>> requests, because if the page would usually respond to POST requests,  
>>> it can only return useful results like Content-length if the HEAD  
>>> request
>>> would upload the same data as POST.
>>
>> I do not understand what you mean here. GET/HEAD are semantically
>> equivalent, not HEAD/POST.
>
> Yes, sure. But a HEAD request is usually being used to obtain a web  
> server's headers on a specific resource. If that resource is configured  
> to respond to POST requests only (e.g. a PHP page responding to uploaded  
> form data), the headers returned by this resource ought to be the same  
> headers a HEAD request would obtain.

No: http://tools.ietf.org/html/rfc2616#section-9.4


> So if a web server's resource would respond to a HTTP POST request with  
> headers containing, e.g., Content-length information (or any header  
> computed from the uploaded POST data) then a HEAD request should yield  
> the same headers. Otherwise the HEAD information retrieved would be  
> useless in order to take decisions on whether to load the full resource  
> using a POST request.
>
> Do you see my point?

I think so, but I believe you are wrong.


-- 
Anne van Kesteren
http://annevankesteren.nl/

Received on Tuesday, 23 February 2010 11:36:37 UTC