Re: [access-control] non-GET threat model and authorization choreography

On Tue, 16 Oct 2007, Henri Sivonen wrote:
> >
> > We can't use OPTIONS because Apache returns
> > 
> >    Allow: GET,HEAD,POST,OPTIONS,TRACE
> > 
> > ...by default, which would basically mean that out of the box, any 
> > resource that support cross-site GET would automatically support 
> > cross-site POST.
> 
> This could be remedied by using a newly named header in the OPTIONS 
> response (e.g. Method-Allow). As a further benefit, introducing new 
> headers would allow the caching outlined in Anne's message.

I have no objections to new headers.


> > Also, OPTIONS doesn't return a body, which is useful to authors who 
> > want to include the cross-domain rights in XML PIs rather than HTTP 
> > headers.
> 
> Do bad things happen if you do return an entity body in an OPTIONS 
> response?

I can't work out how to return _anything_ from a CGI script with OPTIONS. 
The script doesn't seem to get called (using Apache). I'm probably doing 
something wrong, but, as I noted in my reply to Bjoern, this is a security 
issue -- we want the bar to be so low, that even people who don't really 
know what they're doing will still end up safe.


> Moreover, what's the point of using PIs if you have control over HTTP 
> headers?

I think consistency is key here -- if we allow PIs in one area, they 
should work everywhere. Again, security APIs need to be as obvious and 
unsurprising as possible.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

Received on Wednesday, 17 October 2007 01:14:58 UTC