[Gambas-user] PING > Benoît (ServerSocket)
Ron
ron at ...1740...
Fri Nov 6 09:20:28 CET 2009
Kad Mann wrote:
> On Fri, 2009-11-06 at 08:37 +0100, Ron wrote:
>
>> Kad Mann wrote:
>>
>>> On Thu, 2009-11-05 at 14:45 +0100, Jean-Yves F. Barbier wrote:
>>>
>>>
>>>> Benoît Minisini a écrit :
>>>> ...
>>>>
>>>>
>>>>> But if rebooting your Linux server does not make the application work again, I
>>>>> can't really suspect Gambas, as it does not save anything on the hard disk,
>>>>> especially the network component.
>>>>>
>>>>> Did you test the connection with telnet, or any other non-Gambas program? Did
>>>>> you test another socket server application not made in Gambas?
>>>>>
>>>>> The only possibility I see would be a memory corruption in the network
>>>>> component - it is possible, I didn't write it :-). By reading the source code
>>>>> of the Listen() method, I didn't see anything that could lead to that. But I
>>>>> will check again.
>>>>>
>>>>>
>>>> May be a RAM is corrupted, running a full memtest86+ could reveal that
>>>> (well @ least 2 tests: while physically swapping RAMs between tests in order
>>>> to completely test low addresses of the 1st one)
>>>>
>>>>
>>> Gawd! I've got 8Gig to test. I'll take the advice, just to make sure but
>>> the problem has happened on two machines with the very same symptoms.
>>> Thanks.
>>>
>>>
>>>
>> Just thinking out loud maybe it helps.
>>
>> You talked about a Windows machine at one end, maybe also check its
>> firewall and antivirus settings, if any.
>> I know for example Kaspersky 2010 sits between all the network
>> connections you have, maybe some strange pattern is seen as evil, and
>> the connection gets cut off in the middle. Check logs.
>>
>
> I have the problem on linux and linux alone. Windows has nothing to do
> with it.
>
>
>> If you transfer lots of (other) network data, make sure your
>> switches/pcs/routers's network interfaces etc. are in the correct mode,
>> full duplex/duplex,
>>
>
> Well, that'll be a great help if I have only wireless. Nevertheless my
> network is wired but the network is not the culprit either.
>
>
>> lots of equipment doesn't negotiate mode correctly,
>> resulting in one device in half- and the switch in full duplex for
>> example, it seems to work, but not 100%... you get network errors, and
>> strange behavior.
>>
>
> Also nothing to do with it. The problem occurred first when a Windows
> server was involved, however it has also happened since on a standalone
> linux talking to itself on localhost. The second time I posted about the
> problem I didn't mention Windows because Windows has been completely
> eliminated from the equation.
>
>
>> I couldn't see if you are running on 64 Bits Linux, I have had many
>> occasions of problems at the startup of my project, some times my
>> project started 30 times in a row without any problem and the 31th time
>> bingo! It made me nuts back then... It had to do with memory space
>> Gambas was loaded in, if it was in 64Bit memory space, it was unstable.
>> (those bugs are fixed by Benoit so don't worry)
>>
>
> It has happened on both x86 and x64, and I am using x64 right now.
>
>
>> Just to illustrate that sometimes you oversee something obvious at
>> first, and you point at the wrong code just because you don't know where
>> to look.
>>
>
> Yes, true but I do know where not to look.
>
>
>>
Ok fair enough ...
Overlooked the linux-localhost part, so you're right we cannot point at
Bill.
Did you try to run valgrind tool already like Benoit suggested?
It has proven to be very helpful to catch bugs...
I had to run it several times in the past and although the output can be
massive it's worth the try.
Regards,
Ron_2nd.
More information about the User
mailing list