[Quickfix-users] Re: [Quickfix-developers] About passing test cases provided by openfix.net

[Dale W. <wil...@oc...>]

Hi Rohan,

rohan joel pais wrote:

>  
> Hi,
>   What you are telling is right sir, i am unable to recieve the 
> messages from the OpenFix Test Suite while testing for cases 
> 1 - Field Includes SOH character
> 2 - Invalid First three fields.
>
According to the standard (FIX 4.2 w/ errata, page 30)

    Note: The receiving application should disregard any message that is
    garbled, cannot be parsed or
    fails a data integrity check. Processing of the next valid FIX
    message will cause detection of a
    sequence gap and a Resend Request will be generated. Logic should be
    included in the FIX engine to
    recognize the possible infinite resend loop, which may be
    encountered in this situation.


The errors you describe, including SOH in a non-data field, or 
misordering the first three fields produce messages do not pass basic 
data integrity tests, so sending a reject would violate the standard.
I believe the validation suite is in error in expecting a reject.  After 
all, a reject message must identify the sequece number of the message to 
be rejected and you can't parse the message, you don't really know what 
sequence number to reject, right?

Unfortunately this may not help in your situation, so if you MUST 
generate a reject, you'll probably have to hack the QuickFIX code.  At 
least this is Open Source so you have that option.

<plug> If you decide you need to modify the QuickFIX code and are 
uncomfortable with the idea of doing this yourself, OCI, the company I 
work for, does this kind of work on a contract basis.  Let me know 
privately if you want to explore this option da...@oc... </plug>

Dale

>
> In the above 2 cases, i think quickfix engine is ingnoring the 
> messages, and because of that i think these messages can't make it to 
> the session layer. Since my application is not getting these messages, 
> i am not able to send a reject message or so, as requested by the open 
> fix test suite and so i am failing the tests in these 2 cases.
>
> The Test case "Support third party addressing" is not used in our 
> project(and over that its optional) so i am thinking of ignoring it.
>
> The Test case "Synchronize Sequence Numbers" is also an optional test 
> case, and so i am thinking of ignoring it too.
>
> The Test case "Missing Conditionally required field", i think i can 
> handle it with some efforts from my side.
>
>
> Please give me suggestions if i am wrong.
> Thanks in Advance
> with regards
> rohan
>
>
> On Fri, 23 Sep 2005 Caleb Epstein wrote :
> >On 23 Sep 2005 11:17:21 -0000, rohan joel pais <roh...@re...>
> >wrote:
> > >
> > > I have developed an quickfix application which acts as a server. 
> And now i
> > > am testing this application using www.openfix.net 
> <http://www.openfix.net>
> > > .
> > >
> > > I have tested almost all cases except the cases mentioned below
> > >
> > > 1 - Synchronize Sequence Numbers.
> > > 2 - Support third party addressing.
> > > 3 - Invalid First three fields.
> > > 4 - Field Includes SOH character
> > > 5 - Missing Conditionally required field.
> > >
> > > First i thought the above cases (Except 2) will be handled by the 
> quickfix
> > > engine, but it doesn't handle.
> > >
> >You'll need to provide specifics about the tests and what problems 
> you ran
> >into (logfiles and event logs would be helpful). Certain things can 
> probably
> >be made to work with configuration options, but others may not be do-able
> >because of the design of the QuickFIX library. For example, corrupted
> >messages will never make it into the Session layer.
> >
> >
> >--
> >Caleb Epstein
> >caleb dot epstein at gmail dot com
>

-- 
-----------------------------------------------------
  Dale Wilson, Senior Software Engineer      
  Object Computing, Inc. (OCI)
  http://www.ociweb.com/  http://www.theaceorb.com/
----------------------------------------------------