W3C home > Mailing lists > Public > www-lib@w3.org > April to June 2003

Re: Performing a single Request

From: Stewart Johnson <stewart.johnson@ieee.org>
Date: Wed, 28 May 2003 02:51:21 +0930
To: <nin0@gadjahmada.edu>
Cc: www-lib@w3.org
Message-Id: <200305280251.21910.stewart.johnson@ieee.org>


Hi Nino --

It just so happens I'm working on this right now. Here's what I'm working with 
at the moment (cmments below):

--------- begin code ----------
#include "WWWLib.h"
#include "WWWInit.h"
#include "WWWSSL.h"

int main ()
{
    HTRequest * request = NULL;
    HTAnchor * anchor = NULL;
    HTAssocList * formfields = NULL;
    char uri[] = "https://accounts.internode.on.net/cgi-bin/padsl-usage";
    char arg1[] = "username=XXXX";
    char arg2[] = "password=YYYY";
    formfields = HTAssocList_new();
    HTParseFormInput(formfields, arg1);
    HTParseFormInput(formfields, arg2);

    /* create a new pre-emptive client, doesn't use event loop */
    HTProfile_newPreemptiveClient("Sample Post Client", "0.1");

    /* SSL initialisation */
    HTSSL_protMethod_set (HTSSL_V23);
    HTSSL_verifyDepth_set (2);
    HTSSLhttps_init(NO);

    /* Create a request */
    request = HTRequest_new();

    /* Get an anchor object for the URI */
    anchor = HTAnchor_findAddress(uri);

    /* Post the data and get the result in a chunk */
    HTChunk * result = HTPostFormAnchorToChunk(formfields, anchor, request);

    printf("The result from the server was: \"%s\"\n", HTChunk_data(result));

    HTAssocList_delete(formfields);/* Clean up the form fields */
    HTRequest_delete(request);  /* Destroy the Request object */
    HTSSLhttps_terminate();     /* Unregister SSL app support */
    HTProfile_delete();         /* Terminate HTTP client      */
    return 0;
}

--------- end code ----------

This seems to be heading down the right track, I'm no longer using the event 
loop. I got the idea of using HTProfile_newPreemptiveClient from:

http://www.w3.org/Library/Examples/libapp_4.c

The only problem now is that it's not waiting for the request to complete: the 
request is deleted before it's finished being handled:
[stew@uno cpp]$ g++ -o runme simple.c `libwww-config --cflags` `libwww-config 
--libs` -L/usr/lib -lwwwssl
[stew@uno]$ ./runme
Looking up accounts.internode.on.net
Looking up accounts.internode.on.net
Contacting accounts.internode.on.net
The result from the server was: "(null)"
Interrupted!

Like I said, I'm still working on it! Any ideas appreciated!

Stewart.


On Wednesday 28 May 2003 2:28 am, you wrote:
> Dear Stewart,
> I'm having the same problem too..
>
> ---------- Original Message ----------------------------------
> From: Stewart Johnson <stewart.johnson@ieee.org>
> Date: Mon, 26 May 2003 01:27:08 +0930
>
> >G'day --
> >
> >I've been using the the W3C libwww for a little while now, and I've
> > generally had success. I'm having trouble figuring a new problem out, and
> > I'm looking for some advice. Hopefully someone can find the time to steer
> > me in the right direction! :^)
> >
> >The tool I'm currently writing only ever needs to perform a POST request
> > on a single server. It does this request once every half an hour, so for
> > all intents and purposes the requests happen in isolation. I'm writing
> > this tool in C++.
> >
> >I have everything working okay, but I can't help but think there's an
> > easier (and less verbose) way to do it. I'm initialising the library,
> > creating my HTRequest and calling HTPostFormAnchorToChunk to get my
> > result. That all seems correct, my problem is the next part:
> >
> >I'm entering the eventloop using HTEventList_loop. I've previously
> > registered a terminate_handler callback to catch the HTResponse of my
> > HTRequest and save the status in the HTRequest's context, and then call
> > HTEventList_stopLoop. That seems like a whole bunch of unnecessary
> > effort. In addition, it feels ugly because the terminate_handler is a
> > static function.
> >
> >Is it easy to replace the HTEventList_loop? Is there another way of firing
> > off my single HTRequest and catching the response, preferably all within
> > the one function, rather than using a callback?
> >
> >Any advice is greatly appreciated.
> >
> >TIA,
> >Stewart.
Received on Tuesday, 27 May 2003 13:22:12 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 23 April 2007 18:18:43 GMT