[Gambas-user] Connect hang with latest revision

[Ron]

I just checked, gambas2 didn't/doesn't respect timeout either in this 
case when a hiost cannot be reached.
And reports after 2+ minutes.
So it was always like this, maybe because of the fact that it was 
difficult to implement?

Regards,
Ron.

> Thanks.
> The blocking mode issue is not that bad, since I always want to use 
> non-blocking mode anyway, my project is too big to be interupted by 
> sockets.
>
> About the timeout I thought gambas2 did this correctly, but I'm not 
> sure now, since the ClientSocket example uses a timer to remove Socket 
> if it isn't connected in 10 seconds.
>
> Regards,
> Ron_2nd.
>> Hmmm... well I've taken a quick glance at the source and the
>> documentation (Shock news! Developer looks at documentation!!!) and it
>> looks like it's less simple than we thought.
>>
>> When making a connection, it seems as if Gambas insists on using
>> Blocking mode True internally, there is no storage or remembering of the
>> original Blocking mode, so as far as I can see any setting of Blocking
>> mode prior to making a connect is not going to be useful anyway.  In
>> fact, from reading the documentation of Gambas
>> (http://gambasdoc.org/help/comp/gb.net/socket?v3), it looks like it's
>> expected that all Socket comms will be be in non-Blocking (asynchronous)
>> mode:
>>
>> "This class performs its work asynchronously, so the program will not be
>> stopped while connecting, sending or receiving data. "
>>
>> Also it looks like the code relating to Blocking/non-Blocking is
>> provided by the Stream class which Socket inherits.  I think if we're
>> going to make it possible for the user to specify whether they want
>> Blocking or not, and by implication, set a Timeout value... it will need
>> a little more work than just a tweak to call set_timeout whenever
>> set_blocking is called...
>>
>> Kind regards,
>> Caveat
>>
>> On 03/06/13 21:14, Benoît Minisini wrote:
>>> Le 02/06/2013 17:05, Ron a écrit :
>>>> Bug #1 and Bug #2 are fixed, thanks for that!
>>>>
>>>> The time it takes for Event_Error to trigger is still more than 2 
>>>> minutes.
>>>> I see that in ClientSocket there is a 10sec timer running which cuts
>>>> the socket if still not connected.
>>>>
>>>> Is this the way it should be implemented, what about the TimeOut
>>>> property it seems it doesn't work?
>>>> Would be nice if users know after 10 secs that the socket is unable to
>>>> connect, +2 minutes is too long...
>>>>
>>>> Regards,
>>>> Ron.
>>>>
>>> Hi Ron,
>>>
>>> I have implemented what Caveat told in revision #5696: the timeout is
>>> reset each time the blocking mode changes internally.
>>>
>>> Can you try it and tell me if it changes anything for you?
>>>
>>
>> ------------------------------------------------------------------------------ 
>>
>> How ServiceNow helps IT people transform IT departments:
>> 1. A cloud service to automate IT design, transition and operations
>> 2. Dashboards that offer high-level views of enterprise services
>> 3. A single system of record for all IT processes
>> http://p.sf.net/sfu/servicenow-d2d-j
>> _______________________________________________
>> Gambas-user mailing list
>> Gambas-user at lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/gambas-user
>