[IGSTK-Developers] Loopback test

David Gobbi dgobbi at atamai.com
Mon Mar 13 11:58:41 EST 2006 

Hi Patrick,

I looked at the test log and noticed that it is truncated during the 
"Communication::PrintSelf()":

=======

AC = 2
Logger output saved here:
C:/Patrick/IGSTK/VS71-Debug-Loopback/Testing/Temporary/igstkSerialCommunicationTestLoggerOutput.txt
SerialCommunicationForWindows
SerialCommunicationForWindows (06731750)
  RTTI typeinfo:   class igstk::SerialCommunicationForWindows
  Reference Count: 2
  Modified Time: 1
  Debug: Off
  Observers: 
    none
  TimeoutPeriod: 1000
  ReadTerminationCharacter: 
==========

There's supposed to be a lot more after that :-/


The Communication::PrintSelf() code does something bad right here:

  os << indent << "ReadTerminationCharacter: " << 
m_ReadTerminationCharacter
     << std::endl;

The default value of "char *m_ReadTerminationCharacter" is 255, which is 
also
known as the "end-of-file" character, so printing that character might 
be messing
up the test.

I'll change the "PrintSelf" so that it prints the number value for the 
char, instead
of the char itself.

 - David

P.S. The "special character" issue is also a potential problem for the 
loggers.
Right now, the loggers will log exactly what is sent to them, including 
special
characters... it might be best if the logger replaces special characters 
with
appropriate escape sequences, e.g. a "carriage return" sent to the 
logger would
be escaped as "\r".


Patrick Cheng wrote:
> Hi David,
>
> I guess the problem is from my RS232 female-female plug. The numbering 
> of the pins is a little bit different. Now I figure it out and short 
> the right pin 2&3. it passed the hyper terminal test.
>
> I just ran a quick night test on it. and the 
> igstkSerialCommunicationTest is still failing.
>
> http://public.kitware.com/IGSTK/Sites/build02.isis/Win32-IGSTK-VS71-Debug-Loopback/20060313-0430-Nightly/Test.html 
>
>
> Patrick
>
> David Gobbi wrote:
>> Hi Luis,
>>
>> The loopback that I have working on "beck" just has pins 2-3 shorted, 
>> it doesn't have any of the handshaking pins connected like you have 
>> for zion.  It works for me under both Linux and Windows.
>> You really shouldn't be testing it by doing a "cat" to the 
>> /dev/ttyS0, because the state of the device is unknown unless you 
>> previously set its state using "tcsetattr".  It is much better to 
>> test it from within a terminal program, or from within a C program 
>> where you have used tcsetattr to put the serial port into a known 
>> state.  For instance, I think you see the buffering on your serial 
>> port because your serial port is set to "canonical" mode by default.  
>> In canonical mode, it always buffers until until it receives a 
>> newline (and I suspect that the buffering you are seeing is on the 
>> input, not on the output -- the output buffer is only used when 
>> characters are sent to the serial port too fast for the baud rate to 
>> keep up).  The igstk::SerialCommunication class turns off canonical 
>> mode, so that characters can be received one-at-a-time instead of one 
>> line at a time.
>>
>> Patrick: if you are using a Windows machine, then hyperterm is the 
>> right way to test the loopback.  Luis is right that, if you aren't 
>> seeing anything, you should try turning off flow control.  Also, to 
>> make sure that commands are getting out to the serial port, you can 
>> unplug the loopback and plug in an AURORA or POLARIS.  Then, if you 
>> do "INIT "<carriage return> "BEEP 2"<carriage return> it should beep 
>> for you.  If commands can get out to the AURORA but not back in 
>> through the loopback, you know the loopback has problems.
>>
>> On my side, I'll make another loopback with the extra pins connected 
>> so that I can see what kind of different behavior it has compared to 
>> my single-wire loopback.
>>
>> - David
>>
>> Luis Ibanez wrote:
>>> Hi Patrick,
>>>
>>> Here is the wiring that I'm using in
>>> the Loopback device installed in Zion:
>>>
>>> a)  1====4====6
>>> b)  2====3
>>> c)  7====8
>>>
>>> This looks the same as the one that you posted
>>> in your email.
>>>
>>> In  Linux we tested the device by just opening
>>> two terminals and doing:
>>>
>>>      a)   cat  /dev/ttyS0  (in one terminal)
>>>      b)   ls -R  /  > /dev/ttyS0  (on the other terminal)
>>>
>>> You should see text appearing in the first terminal.
>>>
>>>
>>> Notice that since the serial port is buffered, you may
>>> not see any output if you just send a couple of words.
>>> You may have to send more that some Kb of text to make
>>> sure that the serial port buffer if flushed and you
>>> actually see something in the other side.
>>>
>>> This being said. The IGSTK test is failing in Zion,
>>> and we have never been able to connect to the Polaris
>>> from the Linux machine.
>>>
>>> I just tested the same device on my Windows laptop,
>>> using the Hyperterminal test that you posted, and it
>>> works fine for me. That is, when the loopback device
>>> is connected, I see the text that I'm typing, and
>>> when the device is remove, I don't see the text that
>>> I'm typing.
>>>
>>> Did you selected "Flow control" to "None"  ??
>>>
>>>
>>>
>>>    Luis
>>>
>>>
>>>
>>> ---------------------
>>> Patrick Cheng wrote:
>>>> Hi everyone,
>>>>
>>>> I just made a loopback device and plug it into the windows machine 
>>>> here, and it turned out that the SerialCommunicationTest failed too.
>>>> http://public.kitware.com/IGSTK/Sites/build02.isis/Win32-IGSTK-VS71-Debug-Loopback/20060310-0430-Nightly/Test.html 
>>>>
>>>>
>>>> I followed the wiring instruction in the attached picture. and I 
>>>> try to use a windows program to test if this loopback device is 
>>>> working.
>>>>
>>>> Start -> All Programs -> Accessories -> Communications -> Hyper 
>>>> Terminal
>>>>
>>>> I followed the test under this link:
>>>> http://zone.ni.com/devzone/conceptd.nsf/webmain/D48BA43C59B60277862569EE000501FD 
>>>>
>>>>
>>>> Quote, "You will see the terminal window as shown below. If your 
>>>> serial port is working and the loopback cable is properly 
>>>> connected, the text you type appears in the window. If you remove 
>>>> the cable and type text, it should not appear in this window. If 
>>>> this test behaves differently than described here, see the 
>>>> troubleshooting advice at the end of this section. "
>>>>
>>>> I don't see the text when I am typing, so I guess the loopback is 
>>>> not being set up correctly, I don't know if this is the same 
>>>> situation on Zion-Loopback machine.
>>>>
>>>> Am I doing the wring right? or it's a contact miss problem?
>>>>
>>>> Patrick
>>>>
>>>> ------------------------------------------------------------------------ 
>>>>
>>>>
>>>>
>>>> ------------------------------------------------------------------------ 
>>>>
>>>>
>>>> _______________________________________________
>>>> IGSTK-Developers mailing list
>>>> IGSTK-Developers at public.kitware.com
>>>> http://public.kitware.com/cgi-bin/mailman/listinfo/igstk-developers
>>>
>>> _______________________________________________
>>> IGSTK-Developers mailing list
>>> IGSTK-Developers at public.kitware.com
>>> http://public.kitware.com/cgi-bin/mailman/listinfo/igstk-developers
>>>
>>
>>
>> ------------------------------------------------------------------------
>>
>

More information about the IGSTK-Developers mailing list