Re: Novel HTTP methods and the standard Sun Java SDK (was RE: Revised draft of CBD)

Patrick.Stickler@nokia.com wrote:

> I was not referring to the creation of a servlet responding
> to URIQA requests, but for using the Java API to *make*
> requests to other URIQA-englightened servers.
> 
> If you try to set the request method to anything other than
> one of the "standard" methods, an exception is thrown, but
> *without* the side effect, so you can't just catch and 
> disregard the exception (meaing "I know what I'm doing")
> and proceed.
> 
> I tried to get a simple change made to the SDK where a
> single line of code was moved (not changed) such that the
> method was set prior to the test and thrown exception, so
> that careless or naiive users would still have the "protection"
> of the exception being thrown but savvy (or careless ;-)
> users could disregard the exception and proceed with the
> request, using the specified "non-standard" method.
> 
> The change wouldn't even impact any existing applications,
> which would presume the exception in any case.
> 
> See the attached slides. It's trivially simple.
> 
> At present, you can't use the standard Java SDK for URIQA,
> WebDAV, or any other solution employing "non-standard"
> methods, which is a great pity.

Ah, yeah, the java.net package notoriously sucks. For this and a 
numberous of other reasons, the HTTPClient library was created in Apache 
Jakarta Commons, pretty much anybody serious with HTTP connectivity does 
not use the standard library.

-- 
Stefano.