On 7/14/06, Max Carlson <[EMAIL PROTECTED]> wrote:
HTTPLoader probably should handle making proxied requests internally,
since different runtimes have differing proxy requirements.  It's
essentially just a URL rewrite, correct?  If so, it could be exposed as
shared code to be used by HTTPLoader implementations as needed.  I can
imagine an API that takes a request URL and optional header object and
returns a Laszlo proxy-compliant URL.

This means these methods:

    * get( String:url, Array:headers)
    * post( String:url, String:postbody, Object:headers)
    * put( String:url, String:postbody, Object:headers)
    * delete( String:url, Array:headers)

all need a proxied flag. 


Or else the proxied flag could be a state variable in the HTTPLoader object, set with
a setProxied(boolean) method?

 Also, did we decide on a better name for
delete, since it's a reserved word?  Whatever we decide, all four method
names should be consistent, e.g. doGet() or getRequest().  Any more
thoughts on this?

I guess doGet() etc. would be best. 

-Max

Henry Minsky wrote:
> I'm wondering if the HTTPLoader API should handle the 'proxied' flag
> internally?
> I'm asking because looking at the implementation of
> LzHTTPDatasource.doRequest()  , it hands off a request
> to a loader object, and that request can be proxied or direct. If we
> want to make the loader object
> opaque at this point, it needs to be able to handle building and
> sending direct or proxied requests internally.
>
>
>
> --
> Henry Minsky
> Software Architect
> [EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]>
>



--
Henry Minsky
Software Architect
[EMAIL PROTECTED]

_______________________________________________
Laszlo-dev mailing list
[email protected]
http://www.openlaszlo.org/mailman/listinfo/laszlo-dev

Reply via email to