Comments on: Issues with bad implementations of “Accept” HTTP header http://lab.pheromone.ca/2009/11/16/http-accept-header-implementatio/ Inspiration, Experimentation, Innovation Mon, 21 Nov 2011 20:30:32 +0000 hourly 1 http://wordpress.org/?v=3.0.1 By: karl http://lab.pheromone.ca/2009/11/16/http-accept-header-implementatio/comment-page-1/#comment-233 karl Thu, 26 Nov 2009 20:52:10 +0000 http://lab.pheromone.ca/?p=182#comment-233 It would not work, because you do not know if it is an img tag which requests the resource or an iframe for example. :) It would not work, because you do not know if it is an img tag which requests the resource or an iframe for example. :)

]]> By: Olivier Théreaux http://lab.pheromone.ca/2009/11/16/http-accept-header-implementatio/comment-page-1/#comment-231 Olivier Théreaux Thu, 26 Nov 2009 20:41:36 +0000 http://lab.pheromone.ca/?p=182#comment-231 Your problem is not that your solution would not work, but that the (rails) implementation is a little poor. The little experience I've had with content-negotiation tells me you do NOT want to rely on regular expressions (at least not simplistic ones) to parse Accept headers. If you take the time to properly parse the accept headers, then you can implement an algorithm that could send a 406 if the Accept headers neither include image/ nor */*. Your problem is not that your solution would not work, but that the (rails) implementation is a little poor. The little experience I’ve had with content-negotiation tells me you do NOT want to rely on regular expressions (at least not simplistic ones) to parse Accept headers.

If you take the time to properly parse the accept headers, then you can implement an algorithm that could send a 406 if the Accept headers neither include image/ nor */*.

]]>