[
Lists Home |
Date Index |
Thread Index
]
Francis Norton wrote:
>
> Paul Prescod wrote:
>
> >Nevertheless, there is no reason that the same should not work for
> >input. VS.net merely has a bug that does not permit that. If one uses
> >HTTP POST there is no apparent reason that anything that can be done on
> >OUTPUT should be restricted on INPUT. It seems quite simple to me:
> >
> Two problems with this - first, the the bug appears to extend to the
> implementation of mime for input - at least, I can telnet a
> application/x-www-form-urlencoded POST to my VS.Net application, but not
> a text/xml one. Could be operator malfunction, of course, and I agree
> that the WSDL note appears to permit this idiom.
I didn't mean to imply otherwise. The term "bug" was correct from the
point of view of the spec, but it probably not really an implementation
bug. Clearly they've decided to support some parts of the WSDL/HTTP
binding and not others. Perhaps it was one of those not-enough-time
issues.
> .... But secondly, the the
> parameters for my transaction are in fact for a database query, which
> HTTP 1.1 appears to say [1] is an incorrect use of POST.
Although this would technically work, I agree that this is a misuse of
POST. But no less a misuse if you wrap it in SOAP:Envelope elements! The
bottom of this document addresses database queries directly:
http://www.prescod.net/rest/rpc_for_get.html
Paul Prescod
|
