quickfix-developers — QuickFIX developers forum

Re: [Quickfix-developers] Generating two Test requests in a row

[Oren M. <or...@qu...>]

Are you saying you need to send out two test requests whenever QuickFIX=20=

initiates one, or you need to respond twice when you receive a test=20
request from iLink, or both?

--oren

On Mar 1, 2005, at 11:34 AM, Shane Ryan wrote:

> Has anyone gone through the CME iLink 2.0 certification process=20
> before, and if so have they figured out how to make quickfix generate=20=

> two TestRequest messages in a row?
>
> =A0
>
> I know iLinks behaviour is not inline with the FIX 4.2 specs but I=20
> don=92t think I=92ll be able to get them to change it.
>
> =A0
>
> If anyone knows a quick patch, I can do to make it work as per the=20
> iLink specs it would be appreciated.
>
>  =A0
>
> =A0
>
> Thanks,
>
> =A0
>
> Shane

Re: [Quickfix-developers] Validation of required fields in required repeating group

[Joerg T. <Joe...@ma...>]

Yihu Fang wrote:
> It seems that QuickFIX does not validate any required fields in a 
> repeating group.
> [...] 
> 
> I made the following change in DataDictionary class to allow validation 
> of required repeating group. A boolean argument is passed to addXMLGroup 
> method to notify whether the group is a required group, then the map 
> m_requiredFields is correctly populated by addRequiredField() - similar 
> to components. It also needs to recursively check all fields including 
> those in repeating group in the method checkHasRequired(), instead of 
> using fieldMap.isSetField, a recursive method is called.
> 
> Attached please find the patches for DataDictionary.cpp and 
> DataDictionary.h for QuickFIX release 1.9.4. A file with a few test 
> messages is also included.

Thanks, Yihu. I added this entry to the BugTracker:

http://www.quickfixengine.org/bugtracker/bug.php?op=show&bugid=65

and attached your files to the bug. Please update if neccessary.

Cheers, Jörg

-- 
Joerg Thoennes
                                http://macd.com
Tel.: +49 (0)241 44597-24      Macdonald Associates GmbH
Fax : +49 (0)241 44597-10      Lothringer Str. 52, D-52070 Aachen

[Quickfix-developers] Validation of required fields in required repeating group

[Yihu F. <Yih...@re...>]
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Re: [Quickfix-developers] CharField problem (VC6)

[Alexey Z. <ale...@in...>]

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
  <meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
  <title></title>
</head>
<body bgcolor="#ffffff" text="#000066">
Caleb,<br>
<br>
I use the last (1.9.4) version of cause.<br>
Take a look to the sources - FieldsConvertors.h.<br>
I don't claim this problem as a bug, BUT it's better to send an empty
value than<br>
a broken message.<br>
<br>
<pre class="moz-signature" cols="60">Regards,
		Alexey Zubko
</pre>
<br>
<br>
Caleb Epstein wrote:
<blockquote cite="mid...@ma..."
 type="cite">
  <pre wrap="">On Wed, 02 Mar 2005 09:48:32 -0500, Alexey Zubko
<a class="moz-txt-link-rfc2396E" href="mailto:ale...@in...">&lt;ale...@in...&gt;</a> wrote:

  </pre>
  <blockquote type="cite">
    <pre wrap=""> I've got the following problem:
 
 I put a char field (Side) initiated with '\0' by mistake and got FIX
message truncated.
 I think in this case QF has to check the value in CharConvertor::convert
for 0.
 
 struct CharConvertor
 {
   static std::string convert( char value )
   {
     return std::string( 1, value );
   }
 .........
 
 The following string is not empty and has size 1: 
 std::string    s( 1, '\0');
    </pre>
  </blockquote>
  <pre wrap=""><!---->
Which version of QuickFIX?  There were some issues in QF up to and
including version 1.9.2 which prevented messages with embedded null
bytes from being serialized and deserialized properly.  These should
all be fixed as of 1.9.3 and later (though your Side field would still
be invalid).

  </pre>
</blockquote>
</body>
</html>

Re: [Quickfix-developers] CharField problem (VC6)

[Caleb E. <cal...@gm...>]

On Wed, 02 Mar 2005 09:48:32 -0500, Alexey Zubko
<ale...@in...> wrote:

>  I've got the following problem:
>  
>  I put a char field (Side) initiated with '\0' by mistake and got FIX
> message truncated.
>  I think in this case QF has to check the value in CharConvertor::convert
> for 0.
>  
>  struct CharConvertor
>  {
>    static std::string convert( char value )
>    {
>      return std::string( 1, value );
>    }
>  .........
>  
>  The following string is not empty and has size 1: 
>  std::string    s( 1, '\0');

Which version of QuickFIX?  There were some issues in QF up to and
including version 1.9.2 which prevented messages with embedded null
bytes from being serialized and deserialized properly.  These should
all be fixed as of 1.9.3 and later (though your Side field would still
be invalid).

-- 
Caleb Epstein
caleb dot epstein at gmail dot com

[Quickfix-developers] memory leaks in SocketInitiator::doConnect

[Angel P. <ap...@es...>]

Hi,

            I'm having memory leaks in QF 1.9.4 engine using VC++60 and
Windows XP.

 

SocketInitiator ::doConnect allocates memory for a new SocketConnection
object, this memory is release in SocketInitiator ::doDisconnect but this
function is never called.

 

 

            Someone can help?

 

Thanks in advance.

 

Angel Pinto Lopez

            

 

[Quickfix-developers] CharField problem (VC6)

[Alexey Z. <ale...@in...>]

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
  <meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000066">
Hello,<br>
<br>
I've got the following problem:<br>
<br>
I put a char field (Side) initiated with '\0' by mistake and got FIX
message truncated.<br>
I think in this case QF has to check the value in
CharConvertor::convert for 0.<br>
<br>
struct CharConvertor<br>
{<br>
&nbsp; static std::string convert( char value )<br>
&nbsp; {<br>
&nbsp;&nbsp;&nbsp; return std::string( 1, value );<br>
&nbsp; }<br>
.........<br>
<br>
The following string is not empty and has size 1: <br>
std::string&nbsp;&nbsp;&nbsp; s( 1, '\0');<br>
<br>
<pre class="moz-signature" cols="60">-- 

Regards,
		Alexey
</pre>
</body>
</html>

[Quickfix-developers] Generating two Test requests in a row

[Shane R. <Sr...@ty...>]

Has anyone gone through the CME iLink 2.0 certification process before,
and if so have they figured out how to make quickfix generate two
TestRequest messages in a row?

=20

I know iLinks behaviour is not inline with the FIX 4.2 specs but I don't
think I'll be able to get them to change it.

=20

If anyone knows a quick patch, I can do to make it work as per the iLink
specs it would be appreciated.=20

=20

=20

Thanks,

=20

Shane

RE: [Quickfix-developers] Deadlock when call SocketInitiator.stop ()

[Angel P. <ap...@es...>]

Hi J=F6rg,

In the function=20
void Initiator::stop()=20
{ QF_STACK_PUSH(Initiator::stop)=20

  SessionIDs connected =3D m_connected;
  SessionIDs::iterator i =3D connected.begin();
  for ( ; i !=3D connected.end(); ++i )
    Session::lookupSession(*i)->logout();
 =20
  onStop();
  if( m_threadid )
    thread_join( m_threadid );=20
  m_threadid =3D 0;

  QF_STACK_POP
}

In the line "thread_join( m_threadid );" take place a
WaitForSingleObject(thread).

I have a function in my code call "Desconectar" this function is:

1 bool CConexionFix::Desconectar()
2 {
3  try
4  {
5    m_initiatorFix->stop();
6    m_bDesconectando =3D false;
7  }
8  catch (FIX::ConfigError e)
9  {
10    CString strError;
11    strError.Format("Excepci=F3n FIX %s", e.what());
12    CTrazas::EscribirExcepcion(strError);
13    return false;
14  }
15  return true;
16}

Only when the socket is received data the line 5 produce a deadlock. =
This is
easy of reproduce, for example when I'm receive all securities.

Thank you

Angel Pinto Lopez


-----Mensaje original-----
De: Joerg Thoennes [mailto:Joe...@ma...]=20
Enviado el: martes, 01 de marzo de 2005 15:46
Para: Angel Pinto
CC: qui...@li...
Asunto: Re: [Quickfix-developers] Deadlock when call =
SocketInitiator.stop()

Hi Angel,

>             When I call to function SocketInitiator.stop() and the =
socket
is
> received data my application lock in this point.

There have been some patches in CVS since 1.9.4 which deal with =
stopping
initiators.
Perhaps you could try a CVS snapshot?

>             I'm using VC++6.0 under WXP with version 1.9.4 of QF.

Could you provide more information, e.g. stack trace etc.?

Cheers, J=F6rg

--=20
Joerg Thoennes
                                http://macd.com
Tel.: +49 (0)241 44597-24      Macdonald Associates GmbH
Fax : +49 (0)241 44597-10      Lothringer Str. 52, D-52070 Aachen

Re: [Quickfix-developers] Deadlock when call SocketInitiator.stop()

[Joerg T. <Joe...@ma...>]

Hi Angel,

>             When I call to function SocketInitiator.stop() and the socket is
> received data my application lock in this point.

There have been some patches in CVS since 1.9.4 which deal with stopping initiators.
Perhaps you could try a CVS snapshot?

>             I'm using VC++6.0 under WXP with version 1.9.4 of QF.

Could you provide more information, e.g. stack trace etc.?

Cheers, Jörg

-- 
Joerg Thoennes
                                http://macd.com
Tel.: +49 (0)241 44597-24      Macdonald Associates GmbH
Fax : +49 (0)241 44597-10      Lothringer Str. 52, D-52070 Aachen

Re: [Quickfix-developers] Building outside of source directory

[Joerg T. <Joe...@ma...>]

Caleb Epstein wrote:
> On Tue, 01 Mar 2005 11:15:44 +0100, Joerg Thoennes
> <Joe...@ma...> wrote:
> 
>>You could also use 3 copies to build it for 3 systems or just do something like
>>Actually I do not understand why you need to do it this very special way.
>
> Several reasons come to mind:
> 
> * Time.  If the code can be compiled in an external directory, Bill
> could kick off all three architectures to build in parallel.
> * Convention.  Most GNU tools can be built this way.  GCC specifically
> comes to mind.
> * Separation.  Its much easier to cleanup the build artifacts and be
> sure that your source code is pristine and unmodified if you build in
> a separate directory.  You can even make the sources read-only.

Agreed. The posted reminded me vaguely of the GNU software build feature and I quickly 
browsed the autotools, but did not find any pointers how to implement this in QF. So I 
just asked why Robert does need this. You gave the more constructive answer. Thanks.

> None of these are earth-shattering, but it might be nice if QuickFIX
> supported this.  I'm another person who builds QF on multiple
> architectures (only two) and frequently test patches to the code so I
> frequently need to reconfigure and recompile in the same source tree. 
> I would like the ability to have a separate build dir to make this a
> simpler job (and one that didn't require re-configure'ing when
> changing architectures).

Also spoke to Robert about this. We just need some GNU tool works this way and check in 
which place the QF build setup differs. Any suggestions?

Cheers, Jörg

-- 
Joerg Thoennes
                                http://macd.com
Tel.: +49 (0)241 44597-24      Macdonald Associates GmbH
Fax : +49 (0)241 44597-10      Lothringer Str. 52, D-52070 Aachen

Re: [Quickfix-developers] Building outside of source directory

[Caleb E. <cal...@gm...>]

On Tue, 01 Mar 2005 11:15:44 +0100, Joerg Thoennes
<Joe...@ma...> wrote:
> You could also use 3 copies to build it for 3 systems or just do something like
> Actually I do not understand why you need to do it this very special way.

Several reasons come to mind:

* Time.  If the code can be compiled in an external directory, Bill
could kick off all three architectures to build in parallel.
* Convention.  Most GNU tools can be built this way.  GCC specifically
comes to mind.
* Separation.  Its much easier to cleanup the build artifacts and be
sure that your source code is pristine and unmodified if you build in
a separate directory.  You can even make the sources read-only.

None of these are earth-shattering, but it might be nice if QuickFIX
supported this.  I'm another person who builds QF on multiple
architectures (only two) and frequently test patches to the code so I
frequently need to reconfigure and recompile in the same source tree. 
I would like the ability to have a separate build dir to make this a
simpler job (and one that didn't require re-configure'ing when
changing architectures).

-- 
Caleb Epstein
caleb dot epstein at gmail dot com

[Quickfix-developers] Deadlock when call SocketInitiator.stop()

[Angel P. <ap...@es...>]

Hi,

 

            When I call to function SocketInitiator.stop() and the socket is
received data my application lock in this point.

 

            I'm using VC++6.0 under WXP with version 1.9.4 of QF.

 

Thank you

 

Re: [Quickfix-developers] Building outside of source directory

[Joerg T. <Joe...@ma...>]

Hi Bill,

> I'm trying to build outside of the source directory, but it fails with some
> compile errors.
> 
> What I do is:
>   cd ../built
>   ../quickfix/configure
>   	I had to delete config.status, otherwise configure failed.
>   make
> 	Fails in FieldBaseTestCase.h:25:30: CPPTest/TestCase.h: No such file
> or directory
> 
> Has anyone succeed in doing such a build?
> 
> (BTW, the reason is that I need to build QuickFIX for three OS's.)

You could also use 3 copies to build it for 3 systems or just do something like

	./configure --prefix ....
	make
	...
	make install
	make distclean

	# other OS
	./configure --prefix ...
	etc. etc.

Actually I do not understand why you need to do it this very special way.

Cheers, Jörg

-- 
Joerg Thoennes
                                http://macd.com
Tel.: +49 (0)241 44597-24      Macdonald Associates GmbH
Fax : +49 (0)241 44597-10      Lothringer Str. 52, D-52070 Aachen

[Quickfix-developers] Building outside of source directory

[Bill R. Hr. <bil...@ra...>]

I'm trying to build outside of the source directory, but it fails with some
compile errors.

What I do is:
  cd ../built
  ../quickfix/configure
  	I had to delete config.status, otherwise configure failed.
  make
	Fails in FieldBaseTestCase.h:25:30: CPPTest/TestCase.h: No such file
or directory

Has anyone succeed in doing such a build?

(BTW, the reason is that I need to build QuickFIX for three OS's.)
  
Best regards
Robert

Re: [Quickfix-developers] [CVS] configure and _XOPEN_SOURCE

[Caleb E. <cal...@gm...>]

On Mon, 28 Feb 2005 14:24:42 -0600, Oren Miller <or...@qu...> wrote:

> Well, this used to be set to 1, but Bud Griffin reported that Solaris
> needs it to be 500:
> http://sourceforge.net/mailarchive/message.php?msg_id=10522888

It probably depends on the compiler and possibly operating system
release.  That message doesn't specify which compiler or Solaris
version he had trouble with.  We use SunPRO 5.3 on Solaris 2.8.  He
may be using gcc, or perhaps a less-buggy Sun C++ compiler.  CC -V
reports here:

CC: Sun WorkShop 6 update 2 C++ 5.3 Patch 111685-20 2004/03/19

The compiler can be detected by checking for the preprocessor macro
__SUNPRO_CC <= 0x530 if this helps.

Reading through the various system headers (<time.h> and
<sys/feature_tests.h>), setting _XOPEN_SOURCE=4 should be sufficient,
but this setting generates similar errors at compile time.

My best-guess solution for this would be to NOT define _XOPEN_SOURCE
if __SUNPRO_CC <= 0x530 as it doesn't seem to be required.  The
strptime prototype seems to be pulled in just fine without it.

This is also causing a cascading error effect.  Because compiler
errors are generated with _XOPEN_SOURCE defined, the configure test
for socklen_t ends up failing to compile.  So configure thinks it
needs to define the macro socklen_t -> int, which causes additional
errors.

-- 
Caleb Epstein
caleb dot epstein at gmail dot com

Re: [Quickfix-developers] [CVS] configure and _XOPEN_SOURCE

[Oren M. <or...@qu...>]

Well, this used to be set to 1, but Bud Griffin reported that Solaris 
needs it to be 500: 
http://sourceforge.net/mailarchive/message.php?msg_id=10522888

???

--oren

On Feb 28, 2005, at 2:08 PM, Caleb Epstein wrote:

> QuickFIX Documentation: 
> http://www.quickfixengine.org/quickfix/doc/html/index.html
> QuickFIX FAQ: 
> http://www.quickfixengine.org/wikifix/index.php?QuickFixFAQ
> QuickFIX Support: http://www.quickfixengine.org/services.html
>
> This line in configure.in:
>
> AC_DEFINE(_XOPEN_SOURCE, 500, "Define _XOPEN_SOURCE to make sure we
> get strptime")
>
> Causes problems building QF on Solaris.  When _XOPEN_SOURCE is defined
> to 500 in config.h, I get this error compiling any code:
>
> "/usr/include/sys/siginfo.h", line 248: Error: Type name expected
> instead of "clock_t".
> "/usr/include/sys/siginfo.h", line 250: Error: Type name expected
> instead of "clock_t".
>
> Omitting this macro removes the problem.  It looks like a bug in
> Solaris, but I can't figure out how else to fix this.
>
> -- 
> Caleb Epstein
> caleb dot epstein at gmail dot com
>
>
> -------------------------------------------------------
> SF email is sponsored by - The IT Product Guide
> Read honest & candid reviews on hundreds of IT Products from real 
> users.
> Discover which products truly live up to the hype. Start reading now.
> http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
> _______________________________________________
> Quickfix-developers mailing list
> Qui...@li...
> https://lists.sourceforge.net/lists/listinfo/quickfix-developers
>

[Quickfix-developers] [CVS] configure and _XOPEN_SOURCE

[Caleb E. <cal...@gm...>]

This line in configure.in:

AC_DEFINE(_XOPEN_SOURCE, 500, "Define _XOPEN_SOURCE to make sure we
get strptime")

Causes problems building QF on Solaris.  When _XOPEN_SOURCE is defined
to 500 in config.h, I get this error compiling any code:

"/usr/include/sys/siginfo.h", line 248: Error: Type name expected
instead of "clock_t".
"/usr/include/sys/siginfo.h", line 250: Error: Type name expected
instead of "clock_t".

Omitting this macro removes the problem.  It looks like a bug in
Solaris, but I can't figure out how else to fix this.

-- 
Caleb Epstein
caleb dot epstein at gmail dot com

[Quickfix-developers] Re: Patch: prevent overlapping resend requests, next/nextQueued stack overflow

[Caleb E. <cal...@gm...>]

On Mon, 28 Feb 2005 11:46:47 -0500, Caleb Epstein
<cal...@gm...> wrote:

> Attached is a patch which addresses the two problems I reported on
> Friday of last week.

Here's another version of the patch which includes fixes for the
broken test cases.  Now all unit and acceptance tests pass.

-- 
Caleb Epstein
caleb dot epstein at gmail dot com

[Quickfix-developers] Patch: prevent overlapping resend requests, next/nextQueued stack overflow

[Caleb E. <cal...@gm...>]

Attached is a patch which addresses the two problems I reported on
Friday of last week.

It unrolls the recusion between Session::next and ::nextQueued by
passing an extra parameter to the one-argument versions of
Session::next that indicates whether the message is coming from
nextQueued.  This changes the public interface to Session (though the
extra argument it defaults) so it may not be the cleanest solution to
the problem.

The multiple overlapping resend request issue is addressed more or
less as I documented in the earlier email.  SessionState keeps track
of the details of any outstanding resend request and this data is
reset when the session disconnects or a queued message is successfully
processed.

After these changes all of the acceptance tests pass but the unit test
called 10_MsgSeqNumGreater fails for each FIX version.  The input file
for that test bears the comment "ERROR, Why are we sending this
twice?".  I think my patch fixes this bug in the engine, so the
testcase needs to be adjusted :-)  I'm attaching the results of the
runat program.

-- 
Caleb Epstein
caleb dot epstein at gmail dot com

[Quickfix-developers] how do you get doxygen to make all that pretty documentation

[Clark S. <cla...@ya...>]

I can't get the source code  in the *.cpp files to get
included. I messed with this for many hours.

http://66.9.177.253/automate.doc.cfg.txt

the final result for this doxygen file
http://66.9.177.253/automate/html/index.html

the problem is I can only get [code] for the header
files and not the html files
http://66.9.177.253/automate/html/files.html

the project I am documenting, for anybody who is
interested. I am documenting a fix client server app.
http://66.9.177.253/index.html




		
__________________________________ 
Do you Yahoo!? 
Yahoo! Mail - now with 250MB free storage. Learn more.
http://info.mail.yahoo.com/mail_250

Re: [Quickfix-developers] [CVS] SocketException::errorToWhat doesn't return anything on non-Win32

[Joerg T. <Joe...@ma...>]

Hi Caleb,

>>From the latest QF CVS code:
> 
> /home/ietdev/tools/linux-i686-gcc/quickfix-CVS-20050225/include/quickfix/Exceptions.h:
> In member function `std::string
> FIX::SocketException::errorToWhat(int)':
> /home/ietdev/tools/linux-i686-gcc/quickfix-CVS-20050225/include/quickfix/Exceptions.h:347:
> warning: no return statement in function returning non-void
> 
> I think this branch of the code should return "strerror (error)"

Yes, it looks like the non-MS branch of this code looks a bit incomplete. IMHO, Oren 
should check this.

Cheers, Jörg

-- 
Joerg Thoennes
                                http://macd.com
Tel.: +49 (0)241 44597-24      Macdonald Associates GmbH
Fax : +49 (0)241 44597-10      Lothringer Str. 52, D-52070 Aachen

[Quickfix-developers] [CVS] SocketException::errorToWhat doesn't return anything on non-Win32

[Caleb E. <cal...@gm...>]

From the latest QF CVS code:

/home/ietdev/tools/linux-i686-gcc/quickfix-CVS-20050225/include/quickfix/Exceptions.h:
In member function `std::string
FIX::SocketException::errorToWhat(int)':
/home/ietdev/tools/linux-i686-gcc/quickfix-CVS-20050225/include/quickfix/Exceptions.h:347:
warning: no return statement in function returning non-void

I think this branch of the code should return "strerror (error)"

-- 
Caleb Epstein
caleb dot epstein at gmail dot com

[Quickfix-developers] about the consulting gig

[Clark S. <cla...@ya...>]

I forgot to add I have extensive experience working
with traders, about 6 months experience with FIX

http://66.9.177.253/resume.html


		
__________________________________ 
Do you Yahoo!? 
Yahoo! Sports - Sign up for Fantasy Baseball. 
http://baseball.fantasysports.yahoo.com/

[Quickfix-developers] Looking for consulting work

[Clark S. <cla...@ya...>]

I will be finishing up my current consulting gig (LOL
assuming I actually finish the work on time) about
April 1.  There has to be some sort of dramatic
forboding, that I estimated the complection date on
April Fools day.

But any way, back to the point of this email. I am
looking for consulting work, that will let me work
from home several days / week. I can travel anywhere
in the US for the other days. Does anybody have any
suggestions?

I posted a resume on Dice.com, and got positive
feedback in less than a day. Does anybody have any
other sites like Dice.com to suggest?


	
		
__________________________________ 
Do you Yahoo!? 
Yahoo! Mail - You care about security. So do we. 
http://promotions.yahoo.com/new_mail