quickfix-developers — QuickFIX developers forum

[Quickfix-developers] Verbose logging/debugging

[James W. <wi...@ra...>]

 Folks,

   I'm working out the details of a new application and
am running into unexplained session disconnects.  The only
output I find in the logs is "Connection Dropped" which is
not terribly informative.  I would like to know two things:

1) Is there a flag or option for the configuration file
that can get QF to produce more verbose output?

2) Does the flag to ignore user-defined fields apply only
to application-level messages but not to session-level or
administrative messages?

many thanks,
Jim Wiggs

Re: [Quickfix-developers] setup as a router

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

It has been, the terminology just isn't standardized. Usually the 
question comes in the form of running an initiator and acceptor in the 
same process. Like this for instance:

http://thread.gmane.org/gmane.comp.finance.quickfix.devel/1380

Brian McAndrews wrote:

>>Can it be done? Yes. We have an application that accepts multiple FIX 
> connections and translates them to a proprietary ECN interface.
>>Doing a FIX-to-FIX version would probably be considerably simpler.
>>Dale
>
> Dale,
>
> Thanks
>
> Yes, a FIX-to-FIX is what I’m looking at. I’m sure that it shouldn’t 
> be too difficult. It just surprised me some that it apparently hasn’t 
> been discussed here.
>
> Brian
>

RE: [Quickfix-developers] setup as a router

[Brian M. <bri...@ro...>]

>Can it be done?  Yes.  We have an application that accepts multiple FIX
connections and translates them to a proprietary ECN interface.
>Doing a FIX-to-FIX version would probably be considerably simpler.
>Dale
=20

Dale,

Thanks

=20

Yes, a FIX-to-FIX is what I'm looking at.  I'm sure that it shouldn't be
too difficult.  It just surprised me some that it apparently hasn't been
discussed here.

=20

Brian

Re: [Quickfix-developers] Need some help regarding order request

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

Hi, V.Kishore  (what's the proper short form of your name?)

>Dear all,
>
>	I am in the initial state of developing the
>application using QuikFix protocol. In my application
>Initiator needs to send an order to Acceptor, and in
>my order I have to send five fields. We have
>successfully established connection between Initiator
>and Acceptor, now I need to send an order via this
>connection, but I am unable to do it. Please assist me
>in transferring the data between Initiator and
>Acceptor through this connection (I am working on JAVA
>platform). 
>	Please also assist me in creating session ID for the
>above order. And I request to please help me in this
>concern as soon as possible.
>
>  
>
To get started with QuickFIX, take a look at the example programs that 
you installed when you installed quickfix.  They are in quickfix/examples.
You should also read the documentation that you can find starting from 
quickfix/doc/html/index.html (load this into your browser).

To address your specific questions: 

To send an order you should create a FIXnn::NewOrderSingle object (nn is 
a protocol version like 42 or 43) and add the necessary fields to it, 
then send it via one of the FIX::Session methods.
Receive the order by overriding the fromApp method of your 
implementation of the FIX::Application interface.  You should probably 
pass the message to a FIX::MessageCracker and catch the result in an 
overidden onMessage method..  The documentation will help with the details.

The session ID is assembled from the BeginString, SenderCompID, 
TargetCompID, and optional SessionQualifier fields from your fix 
configuration file.
BeginStirng is determined by the version of the protocol you are using: 
i.e. FIX.4.2
SenderCompID and TargetCompID are determined by mutual agreement between 
the coparties (but in practice they are usually assigned by the acceptor 
which is often and exchange).
SessionQualifier can be ignored unless you know why you need it.

Dale

FYI: My company offers an "Introduction to QuickFIX for C++ programmers" 
class to help new developers get started.  Let me know if you'd like 
more details.

>With Regards
>V.Kishore Kumar								
>
>
>
>
>Thanks & Regards
>V.Kishore Kumar
>Software Engineer
>GWN Communications Pvt.Ltd
>Mail to:ki...@gw...,
>        kis...@ya...
>Mobile :+91 9866023792
>
>  
>


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

[Quickfix-developers] Need some help regarding order request

[kishorekumar v. <kis...@ya...>]

Dear all,

	I am in the initial state of developing the
application using QuikFix protocol. In my application
Initiator needs to send an order to Acceptor, and in
my order I have to send five fields. We have
successfully established connection between Initiator
and Acceptor, now I need to send an order via this
connection, but I am unable to do it. Please assist me
in transferring the data between Initiator and
Acceptor through this connection (I am working on JAVA
platform). 
	Please also assist me in creating session ID for the
above order. And I request to please help me in this
concern as soon as possible.


With Regards
V.Kishore Kumar								




Thanks & Regards
V.Kishore Kumar
Software Engineer
GWN Communications Pvt.Ltd
Mail to:ki...@gw...,
        kis...@ya...
Mobile :+91 9866023792


		
__________________________________________ 
Yahoo! DSL – Something to write home about. 
Just $16.99/mo. or less. 
dsl.yahoo.com 

Re: [Quickfix-developers] setup as a router

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

Hi Brian,

Brian McAndrews wrote:

> Surprisingly, I didn't find any info on this in the archives.  My 
> question is, can (or how difficult would it be to setup) quickfix be 
> configured as a router? For instance, multi-inputs to one exchange or 
> multi-inputs to multi exchanges.
>
Can it be done?  Yes.  We have an application that accepts multiple FIX 
connections and translates them to a proprietary ECN interface.

How easy was it?   Not very.   It took a lot of work and a month or so 
to get it right.  The problems however, weren't so much in QuickFIX as 
they were in configuring the system and adapting the FIX view of the 
world to that of the proprietary interface.  Among other things we had 
to define some custom fields to push info through that wasn't readily 
available in the standard FIX fields.  We also spent a lot of time 
coping with a difference in opinion between FIX and the other guy 
concerning what to do when the connection dropped unexpectedly. Doing a 
FIX-to-FIX version would probably be considerably simpler.


Dale


>  
>
> Thanks,
>
> Brian
>


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

Re: [Quickfix-developers] setup as a router

[Andrew M. <an...@nm...>]

Hi Brian.  I have done that. I have also created a class that handles 
basic FIX version conversion since my trading workstations all talk 4.1 
and I'm talking to 4.0 and 4.1 exchanges (Inet(4.1), Arca(4.1), Brut(4.0), 
Goldman's RediFIX(4.1)...
Andrew
an...@nm...

On Tue, 6 Dec 2005, Brian McAndrews wrote:

> Surprisingly, I didn't find any info on this in the archives.  My
> question is, can (or how difficult would it be to setup) quickfix be
> configured as a router? For instance, multi-inputs to one exchange or
> multi-inputs to multi exchanges.
>
>
>
> Thanks,
>
> Brian
>
>

[Quickfix-developers] setup as a router

[Brian M. <bri...@ro...>]

Surprisingly, I didn't find any info on this in the archives.  My
question is, can (or how difficult would it be to setup) quickfix be
configured as a router? For instance, multi-inputs to one exchange or
multi-inputs to multi exchanges.

=20

Thanks,

Brian

RE: [Quickfix-developers] QuickFIX/J tips?

[Steve B. <sb...@sm...>]

Hi J=F6rg,

There were a few bugs in parsing fragmented messages that may
have caused some instability in QFJ with slow connections or
very heavily loaded fast connections. The bugs have been fixed
in CVS. At this point, I'm not sure the stability issue is
related to Netty vs. Mina.

Steve Bate
Smart Trade Technologies
Phone: +33 4 42 90 03 97
http://www.smart-trade.net/
=20

> -----Original Message-----
> From: qui...@li...=20
> [mailto:qui...@li...] On=20
> Behalf Of Joerg Thoennes
> Sent: Tuesday, December 06, 2005 4:55 PM
> To: Tom Dilatush
> Cc: QuickFIX Developers
> Subject: Re: [Quickfix-developers] QuickFIX/J tips?
>=20
> QuickFIX Documentation:=20
> http://www.quickfixengine.org/quickfix/doc/html/index.html
> QuickFIX Support: http://www.quickfixengine.org/services.html
>=20
> Hi Tom,
>=20
> > Just starting to use QuickFIX/J for a production project,=20
> and would be=20
> > appreciative of any tips/suggestions/warnings, etc., if=20
> you'd care to=20
> > share.
>=20
> I had the impression, that the network layer is not stable in=20
> all cases. Please see the recent discussions regarding "move=20
> to MINA." In rare cases the engine stalls somewhere in the=20
> network layer, but I did not investigate further. On the=20
> other hand, we also it in production, but with a moderate=20
> throughput and no issues so far.
>=20
> > The project in question is a system that connects to=20
> several brokers,=20
> > to trade in equities and equity options.
>=20
> 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
>=20
>=20
> -------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc. Do you grep=20
> through log files
> for problems?  Stop!  Download the new AJAX search engine that makes
> searching your log files as easy as surfing the  web. =20
> DOWNLOAD SPLUNK!
> http://ads.osdn.com/?ad_id=3D7637&alloc_id=3D16865&op=3Dclick
> _______________________________________________
> Quickfix-developers mailing list
> Qui...@li...
> https://lists.sourceforge.net/lists/listinfo/quickfix-developers

RE: [Quickfix-developers] SessionSettings improvement (I hope!)...

[Steve B. <sb...@sm...>]

Hi Tom,
 
I don't know if I'd call it a scrubbing but the class has had a light
washing recently. You can see the
changes in the CVS HEAD or QFJ_BRANCH_1_0 branch (the basis of the next
release). The property
name and value parsing now allows a wider range of characters. The line
ending check has also been
made more platform independent although it's probably still not as good as
it could be.
 
The SessionSettings is not an interface because we are maintaining API
compatibility with the C++ 
JNI wrapper. It's possible to programatically set individual nondefault
properties in a SessionSettings 
object. There are many examples in the unit tests. There's currently no way
to set individual
default parameters. That can be added without breaking compatibility with
the QFJ JNI code.
 
[Note: I just added the unit tests and implementation for default property
setters (the getters were already 
there). They are in the QFJ 1.0 branch and the HEAD.]
 
Steve Bate
Smart Trade Technologies
Phone: +33 4 42 90 03 97
http://www.smart-trade.net/
 


  _____  

From: qui...@li...
[mailto:qui...@li...] On Behalf Of Tom
Dilatush
Sent: Tuesday, December 06, 2005 4:54 PM
To: Shepheard, Toby (London)
Cc: QuickFIX Developers
Subject: Re: [Quickfix-developers] SessionSettings improvement (I hope!)...


Hi, Toby...

No argument with any of that.  But actually, I think the whole class could
use a good scrubbing.  That odd construct "\r\n".indexOf(ch) I borrowed from
a piece of code just following the one I modified; it's in there more than
once.  And that whole tokenizer is a bit hard to follow.  Also, I'd like to
see the SessionSettings made into an interface, so that other kinds of
session settings could be created.  In my application, I've wrapped it with
a class that lets me set the settings individually and programatically (so I
can fish the settings from a master config file for the entire app), but the
interface to SessionSettings is kinda wierd -- I generate the InputStream to
feed to SessionSettings <smile>.

Re: [Quickfix-developers] QuickFIX/J tips?

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

Hi Tom,

> Just starting to use QuickFIX/J for a production project, and would be 
> appreciative of any tips/suggestions/warnings, etc., if you'd care to 
> share.

I had the impression, that the network layer is not stable in all cases. Please see the 
recent discussions regarding "move to MINA." In rare cases the engine stalls somewhere in 
the network layer, but I did not investigate further. On the other hand, we also it in 
production, but with a moderate throughput and no issues so far.

> The project in question is a system that connects to several brokers, to 
> trade in equities and equity options.

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] SessionSettings improvement (I hope!)...

[Tom D. <to...@di...>]

Hi, Toby...

No argument with any of that.  But actually, I think the whole class 
could use a good scrubbing.  That odd construct "\r\n".indexOf(ch) I 
borrowed from a piece of code just following the one I modified; it's in 
there more than once.  And that whole tokenizer is a bit hard to 
follow.  Also, I'd like to see the SessionSettings made into an 
interface, so that other kinds of session settings could be created.  In 
my application, I've wrapped it with a class that lets me set the 
settings individually and programatically (so I can fish the settings 
from a master config file for the entire app), but the interface to 
SessionSettings is kinda wierd -- I generate the InputStream to feed to 
SessionSettings <smile>.

Tom...



Shepheard, Toby (London) wrote:
> You might find changing isValueCharacter to treat spaces as a valid 
> value a cleaner approach, though I'm not sure if this might cause 
> other problems. Given that the string is trimmed at the end though I 
> think it would be ok.
>  
> A word of warning on using "\r\n" - this is system dependent. You 
> should use the system newline property instead.
>
>     -----Original Message-----
>     *From:* qui...@li...
>     [mailto:qui...@li...] *On
>     Behalf Of *Tom Dilatush
>     *Sent:* 06 December 2005 15:29
>     *To:* QuickFIX Developers
>     *Subject:* [Quickfix-developers] SessionSettings improvement (I
>     hope!)...
>
>     Hi, folks...
>
>     I've just started using QuickFIX/J for a production project.  I
>     had a bit of trouble getting it going, and it turned out that the
>     problem was a data dictionary setting with a file name that
>     included spaces (this is hosted on a Windows system).  This is
>     inconvenient, nothing more, but I fixed the problem with the
>     change shown below, to the getToken() method of
>     SessionSetting.Tokenizer.  The original code is in blue, the new
>     in red.
>
>     Tom...
>
>                 } else if (ch == '=') {
>                     ch = nextCharacter(inputStream); // wrong
>                     skipWhitespace(inputStream);
>                     if (isValueCharacter(ch)) {
>
>                         /* original block of code, failed on file
>     names including spaces
>                         sb.setLength(0);
>                         do {
>                             sb.append(ch);
>                             ch = nextCharacter(inputStream);
>                         } while (isValueCharacter(ch));
>                         return new Token(VALUE_TOKEN, sb.toString());
>                         */
>
>                         /* replacement block */
>                         sb.setLength(0);
>                         do {
>                             sb.append(ch);
>                             ch = nextCharacter(inputStream);
>                         } while ("\r\n".indexOf(ch) == -1);
>                         return new Token(VALUE_TOKEN,
>     sb.toString().trim());
>                         /* end replacement block */
>                     }
>
> ------------------------------------------------------------------------
> If you are not an intended recipient of this e-mail, please notify the 
> sender, delete it and do not read, act upon, print, disclose, copy, 
> retain or redistribute it. Click here 
> <http://www.ml.com/email_terms/>for important additional terms 
> relating to this e-mail.     http://www.ml.com/email_terms/
> ------------------------------------------------------------------------

RE: [Quickfix-developers] QuickFIX/J tips?

[Shepheard, T. (London) <Tob...@ml...>]

Most problems people have seem to be around starting/stopping sessions
and logon/logoff code so I'd focus on getting these working, and having
robust tests for the various possible scenarios. Other interesting areas
worth focussing on are resends and messages out of sequence. I'd
strongly recommend implementing your own test-tool to mimic the systems
you'll be communicating with, at a basic level at least so they can send
canned responses back. The example applications that come with QFJ are a
great starting point for such tools.

I'd also suggest using the latest code from the main trunk on CVS rather
than the beta release that you can get as a zip file, as there have been
a number of updates in the intervening 3-4 months.

Other than that, not much I can think of that's directly QFJ related.
Good luck!

Toby

-----Original Message-----
From: qui...@li...
[mailto:qui...@li...] On Behalf Of
Tom Dilatush
Sent: 06 December 2005 15:31
To: QuickFIX Developers
Subject: [Quickfix-developers] QuickFIX/J tips?


QuickFIX Documentation:
http://www.quickfixengine.org/quickfix/doc/html/index.html
QuickFIX Support: http://www.quickfixengine.org/services.html

Hi, folks...

Just starting to use QuickFIX/J for a production project, and would be=20
appreciative of any tips/suggestions/warnings, etc., if you'd care to
share.

The project in question is a system that connects to several brokers, to

trade in equities and equity options.

Tom...




-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log
files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=3D7637&alloc_id=3D16865&op=3Dclick
_______________________________________________
Quickfix-developers mailing list
Qui...@li...
https://lists.sourceforge.net/lists/listinfo/quickfix-developers
--------------------------------------------------------

If you are not an intended recipient of this e-mail, please notify the =
sender, delete it and do not read, act upon, print, disclose, copy, =
retain or redistribute it. Click here for important additional terms =
relating to this e-mail.     http://www.ml.com/email_terms/
--------------------------------------------------------

RE: [Quickfix-developers] SessionSettings improvement (I hope!)...

[Shepheard, T. (London) <Tob...@ml...>]

You might find changing isValueCharacter to treat spaces as a valid
value a cleaner approach, though I'm not sure if this might cause other
problems. Given that the string is trimmed at the end though I think it
would be ok.
=20
A word of warning on using "\r\n" - this is system dependent. You should
use the system newline property instead.

	-----Original Message-----
	From: qui...@li...
[mailto:qui...@li...] On Behalf Of
Tom Dilatush
	Sent: 06 December 2005 15:29
	To: QuickFIX Developers
	Subject: [Quickfix-developers] SessionSettings improvement (I
hope!)...
=09
=09
	Hi, folks...
=09
	I've just started using QuickFIX/J for a production project.  I
had a bit of trouble getting it going, and it turned out that the
problem was a data dictionary setting with a file name that included
spaces (this is hosted on a Windows system).  This is inconvenient,
nothing more, but I fixed the problem with the change shown below, to
the getToken() method of SessionSetting.Tokenizer.  The original code is
in blue, the new in red.
=09
	Tom...
=09
	            } else if (ch =3D=3D '=3D') {
	                ch =3D nextCharacter(inputStream); // wrong
	                skipWhitespace(inputStream);
	                if (isValueCharacter(ch)) {
=09
	                    /* original block of code, failed on file
names including spaces
	                    sb.setLength(0);
	                    do {
	                        sb.append(ch);
	                        ch =3D nextCharacter(inputStream);
	                    } while (isValueCharacter(ch));
	                    return new Token(VALUE_TOKEN,
sb.toString());
	                    */
=09
	                    /* replacement block */
	                    sb.setLength(0);
	                    do {
	                        sb.append(ch);
	                        ch =3D nextCharacter(inputStream);
	                    } while ("\r\n".indexOf(ch) =3D=3D -1);
	                    return new Token(VALUE_TOKEN,
sb.toString().trim());
	                    /* end replacement block */
	                }
--------------------------------------------------------

If you are not an intended recipient of this e-mail, please notify the =
sender, delete it and do not read, act upon, print, disclose, copy, =
retain or redistribute it. Click here for important additional terms =
relating to this e-mail.     http://www.ml.com/email_terms/
--------------------------------------------------------

[Quickfix-developers] QuickFIX/J tips?

[Tom D. <to...@di...>]

Hi, folks...

Just starting to use QuickFIX/J for a production project, and would be 
appreciative of any tips/suggestions/warnings, etc., if you'd care to share.

The project in question is a system that connects to several brokers, to 
trade in equities and equity options.

Tom...

[Quickfix-developers] SessionSettings improvement (I hope!)...

[Tom D. <to...@di...>]

Hi, folks...

I've just started using QuickFIX/J for a production project.  I had a 
bit of trouble getting it going, and it turned out that the problem was 
a data dictionary setting with a file name that included spaces (this is 
hosted on a Windows system).  This is inconvenient, nothing more, but I 
fixed the problem with the change shown below, to the getToken() method 
of SessionSetting.Tokenizer.  The original code is in blue, the new in red.

Tom...

            } else if (ch == '=') {
                ch = nextCharacter(inputStream); // wrong
                skipWhitespace(inputStream);
                if (isValueCharacter(ch)) {

                    /* original block of code, failed on file names 
including spaces
                    sb.setLength(0);
                    do {
                        sb.append(ch);
                        ch = nextCharacter(inputStream);
                    } while (isValueCharacter(ch));
                    return new Token(VALUE_TOKEN, sb.toString());
                    */

                    /* replacement block */
                    sb.setLength(0);
                    do {
                        sb.append(ch);
                        ch = nextCharacter(inputStream);
                    } while ("\r\n".indexOf(ch) == -1);
                    return new Token(VALUE_TOKEN, sb.toString().trim());
                    /* end replacement block */
                }

Re: [Quickfix-developers] order of tags

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

On 12/6/05, Alvin Wang <AW...@ff...> wrote:
>
> I wonder how the order of tags in a message is determined. It seems that
> it is not the same order that code sets the tags. If that is the case, ho=
w
> can I guarantee that all the tags belong to the same component block (suc=
h
> as Instrument component block) are put together in a message?
>
I don't think there is any requirement that tags of a particular component
be grouped together.  The tag ordering rules as implemented in QuickFIX are
as below:

   - In the header, tags BeginString, BodyLength and MsgType appear first
   and in that order.  Other header fields appear in numeric order.
   - In the body, fields appear in numeric order except for the case of
   repeating groups.  Fields in a repeating group appear in the order defin=
ed
   by the Data Dictionary.
   - In the trailer, CheckSum must appear LAST.  Other fields (Signature,
   SignatureLength) appear in numerical order.

Here's a quote from the FIX 4.4 spec:

Message Format

The general format of a FIX message is a standard header followed by the
message body fields and terminated with a standard trailer.   **

Each message is constructed of a stream of  <tag>=3D<value> fields with a
field delimiter between fields in the stream.  Tags are of data type *TagNu=
m.
* *All tags must have a value specified.  Optional fields without values
should simply not be specified in the FIX message.  A Reject message is the
appropriate response to a tag with no value.*

*Except where noted, fields within a message can be defined in any sequence
(Relative position of a field within a message is inconsequential.) The
exceptions to this rule are:*

*1.       **General message format is composed of the standard header
followed by the body followed by the standard trailer.*

*2.       **The first three fields in the standard header are BeginString
(tag #8) followed by BodyLength (tag #9) followed by MsgType (tag #35).*

*3.       **The last field in the standard trailer is the CheckSum (tag
#10).*

*4.       **Fields within repeating data groups must be specified in the
order that the fields are specified in the message definition within the FI=
X
specification document.  The NoXXX field where XXX is the field being
counted specifies the number of repeating group instances that must
immediately precede the repeating group contents. *

*5.       **A tag number (field) should only appear in a message once. If i=
t
appears more than once in the message it should be considered an error with
the specification document. The error should be pointed out to the FIX
Global Technical Committee.*


--
Caleb Epstein
caleb dot epstein at gmail dot com

Re: [Quickfix-developers] order of tags

[Alvin W. <AW...@FF...>]

Thanks.
Does tags of a component block need to be put together?  I think so...






Dave Linaker <dav...@ma...>
12/05/2005 05:43 PM

 
        To:     Alvin Wang <AW...@FF...>, Oren Miller <or...@qu...>, 
<qui...@li...>, 
<qui...@li...>
        cc: 
        bcc: 
        Subject:        Re: [Quickfix-developers] order of tags


Hi Alvin,

As far as I understand, the default ordering of tags is numeric, i.e. 
lowest tag number first and highest last. The header and trailer are 
slightly different, in that the header must start with tag 8, 9 and 35, 
and the trailer must end with tag 10, but otherwise tags are ordered 
numerically. 

With repeating groups order is more important and hence there is a 
constructor that allows you to specify a message order, e.g.:

    Group( int field, int delim, const int order[] )

...where order[] specifies the tag sort order.

The ordering seems to be implemented in FieldMap, a base class used for 
the various message parts.

Oren, Are there any thoughts on providing a similar constructor in the 
message class?

Dave


On 6/12/05 5:31 am, "Alvin Wang" <AW...@FF...> wrote:

Hi 

I wonder how the order of tags in a message is determined. It seems that 
it is not the same order that code sets the tags. If that is the case, how 
can I guarantee that all the tags belong to the same component block (such 
as Instrument component block) are put together in a message? 

Thanks 

Alvin 
********************************************************************** 
This e-mail message is intended solely for the use of the addressee. The 
message may contain information that is privileged and confidential. 
Disclosure to anyone other than the intended recipient is prohibited. If 
you are not the intended recipient, please do not disseminate, distribute 
or copy this communication, by e-mail or otherwise. Instead, please notify 
us immediately by return e-mail (including the original message with your 
reply) and then delete and discard all copies of the message. We have 
taken precautions to minimize the risk of transmitting software viruses 
but nevertheless advise you to carry out your own virus checks on any 
attachment to this message.  We accept no liability for any loss or damage 
caused by software viruses. 
**********************************************************************

Re: [Quickfix-developers] order of tags

[Dave L. <dav...@ma...>]

Hi Alvin,

As far as I understand, the default ordering of tags is numeric, i.e. lowest
tag number first and highest last. The header and trailer are slightly
different, in that the header must start with tag 8, 9 and 35, and the
trailer must end with tag 10, but otherwise tags are ordered numerically.

With repeating groups order is more important and hence there is a
constructor that allows you to specify a message order, e.g.:

    Group( int field, int delim, const int order[] )

...where order[] specifies the tag sort order.

The ordering seems to be implemented in FieldMap, a base class used for the
various message parts.

Oren, Are there any thoughts on providing a similar constructor in the
message class?

Dave


On 6/12/05 5:31 am, "Alvin Wang" <AW...@FF...> wrote:

> Hi 
> 
> I wonder how the order of tags in a message is determined. It seems that it is
> not the same order that code sets the tags. If that is the case, how can I
> guarantee that all the tags belong to the same component block (such as
> Instrument component block) are put together in a message?
> 
> Thanks 
> 
> Alvin **********************************************************************
> This e-mail message is intended solely for the use of the addressee. The
> message may contain information that is privileged and confidential.
> Disclosure to anyone other than the intended recipient is prohibited. If you
> are not the intended recipient, please do not disseminate, distribute or copy
> this communication, by e-mail or otherwise. Instead, please notify us
> immediately by return e-mail (including the original message with your reply)
> and then delete and discard all copies of the message. We have taken
> precautions to minimize the risk of transmitting software viruses but
> nevertheless advise you to carry out your own virus checks on any attachment
> to this message.  We accept no liability for any loss or damage caused by
> software viruses.
> **********************************************************************

[Quickfix-developers] order of tags

[Alvin W. <AW...@FF...>]

Hi

I wonder how the order of tags in a message is determined. It seems that 
it is not the same order that code sets the tags. If that is the case, how 
can I guarantee that all the tags belong to the same component block (such 
as Instrument component block) are put together in a message?

Thanks
Alvin
                                                                                                                                                                                                                                                                                                    **********************************************************************
This e-mail message is intended solely for the use of the addressee.
The message may contain information that is privileged and confidential.  
Disclosure to anyone other than the intended recipient is
prohibited. If you are not the intended recipient, please do not
disseminate, distribute or copy this communication, by e-mail or
otherwise. Instead, please notify us immediately by return e-mail
(including the original message with your reply) and then delete
and discard all copies of the message. We have taken precautions to
minimize the risk of transmitting software viruses but nevertheless
advise you to carry out your own virus checks on any attachment to
this message.  We accept no liability for any loss or damage caused
by software viruses.
**********************************************************************

[Quickfix-developers] Re: [qfj] move to MINA?

[Chuck C. <chu...@gm...>]

If it's not too much trouble, starting the MINA-related branch sounds
like a good idea to me.  In the long-term, moving Quickfix/J to MINA
seems worthwhile.  

BTW, thanks for the recent check-ins, Steve!

Chuck

RE: [Quickfix-developers] [qfj] move to MINA?

[Steve B. <sb...@sm...>]

Barry Kaplan had been working on a MINA initiator and acceptor. He's
recently started a
new job and is very busy so the work has stalled. We haven't put the code
into the quickfixj
repository because it's still very experimental. However, if some other
people are seriously
interested in working on it (I don't have some at the moment), maybe I could
create a 
MINA-related branch and add the experimental code there for people to work
on.
 
If we create a MINA initiator and acceptor I'd suggest removing the
poll/block methods
from the classic initiator/acceptor pair. My understanding is they are
intended for single
threaded applications. However, no Java application is truly single threaded
and most
networking application are multithreaded in the Java world.
 
What does everyone think about this? I've noticed a few posts that imply
that some
people are using block and/or poll with Java. I'm interested in hearing what
benefit
people are getting from those methods (since they don't really reduce the
number
of threads in pure Java implementation).
 
Steve Bate
Smart Trade Technologies
Phone: +33 4 42 90 03 97
http://www.smart-trade.net/
 


  _____  

From: qui...@li...
[mailto:qui...@li...] On Behalf Of Ian
Johnson
Sent: Thursday, December 01, 2005 5:44 PM
To: qui...@li...
Subject: [Quickfix-developers] [qfj] Hexdumps under Netty2 1.8.0, move to
MINA?


Hi,
 
Definitely a vote for this. My requirement is for better latency management
and embedded SSL - both of which can be done fairly easily in MINA (or so it
says). If you have a MINA version of the Initiatior and FIXMessage I would
definitely be interested.
 
Any change of posting the changes?
 
I was also seeing buffer underflow issues until I patched the parsing code
(I see thats changed recently, but I haven't examnined the changes)
 
My requirements is for latency rather than high throughput (FX arbitrage
where most of what matters is latency)
 
Good work chaps
 
Ian

[Quickfix-developers] (no subject)

[Stanislawek, M. <MSt...@cm...>]

confirm 953176

[Quickfix-developers] JNI Acceptor cannot sustain multiple connections

[Stanislawek, M. <MSt...@cm...>]

We've experienced an upper bound on the number of configured sessions
and connected clients an acceptor can handle through our quickfix Java
JNI implementation.  At first we thought it was file descriptor based,
but after bumping up both the hard and soft limits well above what is
required and no change in behavior, we think it's not the shell's file
descriptor limits were hitting and maybe something jvm related.

We also took our app out of the equation by using the example java
executor that is in the quickfix distribution.  And the same behavior
occurred.  Here are the details:

We modified the java executor to use the file store and file log (this
is what we'd like to use in prod) and modified the application to print
the messages in the fromAdmin and fromApp call backs.  What I notice is
at 144 configured sessions I can establish a connection, login and start
heart beating.  I see the logon and heartbeat messages printed to
stdout.  But at 145 sessions configured, the app callbacks never sees
the logon message.  netstat shows the connection established but the
receive queue is never emptied.  netstat shows the receive queue
containing the ~108 bytes from the logon message.

Has anyone else bumped up against this?

Here is the environment:
RHAS 3.1
quickfix 1.10.2
g++ 3.2.3
Sun's 1.4.2_02 jvm

UseDataDictionary=3DN
ResetOnLogout=3DY
ResetOnDisconnect=3DY

/tmp/executor $ ulimit -a
address space limit (kbytes)   (-M)  unlimited
core file size (blocks)        (-c)  0
cpu time (seconds)             (-t)  unlimited
data size (kbytes)             (-d)  unlimited
file size (blocks)             (-f)  unlimited
locks                          (-L)  unlimited
locked address space (kbytes)  (-l)  4
nofile                         (-n)  65535
nproc                          (-u)  7168
pipe buffer size (bytes)       (-p)  4096
resident set size (kbytes)     (-m)  unlimited
socket buffer size (bytes)     (-b)  4096
stack size (kbytes)            (-s)  10240
threads                        (-T)  not supported
process size (kbytes)          (-v)  unlimited

[Quickfix-developers] [qfj] Hexdumps under Netty2 1.8.0, move to MINA?

[Ian J. <Ian...@bt...>]

Hi,

Definitely a vote for this. My requirement is for better latency =
management and embedded SSL - both of which can be done fairly easily in =
MINA (or so it says). If you have a MINA version of the Initiatior and =
FIXMessage I would definitely be interested.

Any change of posting the changes?

I was also seeing buffer underflow issues until I patched the parsing =
code (I see thats changed recently, but I haven't examnined the changes)

My requirements is for latency rather than high throughput (FX arbitrage =
where most of what matters is latency)

Good work chaps

Ian