get URL returning empty string

[Dave Cragg]

On 2 Jun 2005, at 18:14, Alex Tweedly wrote:

> Dave Beck wrote:
>
>
>> Thanks Chipp!!
>>
>> But unfortunately putting the command in a repeat structure like that
>> doesn't seem to be helping. For some reason "get URL" returning  
>> empty is
>> quite consistent behavior on at least one customer's machine.
>>
>> Might it help to use the "load" command instead? Are they  
>> implemented more
>> or less the same way internally? What about communicating through  
>> sockets?
>> Is that more reliable? (I am currently using a .php script to  
>> execute some
>> commands on the server and return a value.)
>>
>>
>>
> Not directly answering your question, but ....
>
> on the machines which fail, can you try accessing the Web page (php  
> script) from a browser and verify that works.
> I know you said
>
>
>> I know that the web page that I am
>> trying to load is NOT empty.
>>
>>
> but I explicitly want to check for accessibility from the same  
> machine on which it fails via Rev.

After doing what Alex suggests, you should probably use  
libUrlSetLogField to log the exchange between your app and the  
server. The odd thing is that "the result" is returning empty. This  
should only happen if the server sends a 200 response. If it is  
getting a 200 response, I'd want to see whether there is a "Content- 
length" header in the server's reponse, and if not, whether there is  
a Transfer-encoding" header. If there is neither, libUrl will use a  
"socketClosed" message to determine the end of the data transfer. In  
this case, a premature close of the socket might produce what you are  
seeing. (When those headers are present, which is normal, a  
socketClosed message during a data transfer will produce an error in  
the result.)

This is the only thing that I can think of to explain what you are  
seeing, unless it's a bug in libUrl. In either case, getting the  
server's response would help.

Cheers
Dave