>>>>>>Hi,
>>>>>>
>>>>>>Finally got back to doing some more testing of the last code you sent.
>>>>>>Unfortunately, if I set up a connection with an invalid gateway, then the API call to GetHostByName is slow.....
>>>>>>As of now I can't see a way around that one ?
>>>>>>
>>>>>>Best,
>>>>>>Viv
>>>>>

>>>>>

>>>>>
>>>>>Viv,
>>>>>
>>>>>I've made it so that GetHostByName() is only called if you use www.xxxx.xx style
>>>>>If you pass an ip address - it does not call GetHostByName()
>>>>
>>>>Hi,
>>>>I'd seen (and tested) that using a fixed IP address as the 'hit point' would work in a timely fashion. But the code to test for access to the address will be embedded in the application and I cannot guarantee that the IP address currently assigned to the URL will remain static over the medium term. However I do have access to a number of fixed IP addresses so maybe I could use one of those to route the connection ( I could at least guarantee that domain name resolution would work at that point). Need to dwell on this......
>>>>
>>>
>>>haha - I just answered how to get around GetHostByName()
>>>
>>>How many customers with an invalid gateway will you have ?
>>Not many I suppose. But I don't want to have to wait a couple of minutes before telling them........
>>
>>>Is an invalid gateway of any use to you ?
>>Haven't thought of a good use for one yet :-}
>>
>>I've also been testing a few scenarios with the MSSoap com stuff (via the VFP _webservices lib). There's odd behaviour there when there's a valid gateway but an unreachable (or invalid) DNS address. The soap client .MSSoapInit() call may take a long time to complete but eventually reports success. Subsequent calls can succeed almost instantly but of course any attempt to actually use the service fails :-{
>>
>>All I want to be able to do is determine *quickly* whether the VFP application is capable of successfully making a simple web service call to a given URL. Would be a doddle in .NET!
>
>
>Oh, and how would you do it in .net ? Tell me .

Spin whatever it takes off onto another thread? If there's no callback in, say, 10 seconds I'd be happy to assume no connection.

>If you can do it in .net then it is possible with api since I reckon a lot of .net classes are just an interface to api calls
>
>Oh, I think I have an idea . Something for the weekend ..