On Fri, Sep 11, 2009 at 2:05 PM, Paul Davis <[email protected]> wrote: > On Fri, Sep 11, 2009 at 4:56 PM, Zachary Zolton > <[email protected]> wrote: >> AFAICT Paul is correct; it seems to be merely an omission from the >> HTTP spec that GET requests don't explicitly forbid sending an entity >> body. Moreover, if caching proxies don't respect the entity body, it >> wont solve my original problem anyways. >> >> So, it seems the answer would be for my application to better >> normalize my documents for efficient querying. (No surprise!) >> >> Too bad though, that did sound like a fun bite-size patch on which to >> practice my Erlang... :^D >> > > Feel free to submit a patch to add the URL string parameter. Its been > oft requested enough that it should probably go in. I'll get to it > eventually otherwise. > > Paul Davis
This is related to a patch we need: a general JSON syntax for queries. The idea is that instead of POSTing an array of keys to a view, you'd post an array of query definition objects. This way you could do multiple ranges, and other things. > >> On Fri, Sep 11, 2009 at 1:20 PM, Paul Davis <[email protected]> >> wrote: >>>> We could also start supporting GET bodies :) >>> >>> Its not about us supporting GET bodies. Its about everyone else >>> supporting them. And by everyone, I mean everyone from http client >>> libraries to proxy implementations. >>> >>> Last I recall on the subject was that GET bodies only arguably exist >>> because they're never explicitly forbidden. I'm not particularly >>> familiar with all the details though so I could be wrong on that >>> point. >>> >>> Paul Davis >>> >> > -- Chris Anderson http://jchrisa.net http://couch.io
