quickfix-developers — QuickFIX developers forum

Re: [Quickfix-developers] C++ Quickfix Memory keeps growing

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

I would make a change to populate the map from the header file only 
during the resend request.
The binary format for header could speed up the resend process, but I 
don't think it's really matter.

Regards,
	Alexey Zubko




Djalma Rosa dos Santos Filho wrote:
> Yes, this is a pseudo memory leak.
>
> Keeping all outbound sequence numbers in maps wasn't a smart solution 
> for the FileStore.
>
> To avoid development effort, you can use the other available store 
> factories with their pros and cons (memory, MySQL, ODBC etc).
>
> But, what sounds to me as a solution to save the current FileStore 
> from this problem is the .header file in a binary format (a sequence 
> of fixed blocks / serialized structs). In this file QF would append 
> and seek on demmand the seqnum's position data (offset | message 
> length) in corresponding .body file (additionally a cache for the LRU 
> would be nice, but not all data in memory).
>
> On Wed, Jan 21, 2009 at 5:42 PM, Alexey Zubko 
> <ale...@in... 
> <mailto:ale...@in...>> wrote:
>
>     QuickFIX Documentation:
>     http://www.quickfixengine.org/quickfix/doc/html/index.html
>     QuickFIX Support: http://www.quickfixengine.org/services.html
>
>     That's probably the maps keeping messages for resend (store factory).
>     Take a look into the /PersistMessages /configuration parameter.
>     As alternative you can write your own store factory.
>
>     Regards,
>            Alexey Zubko
>
>
>
>
>
>     Michele Chitson wrote:
>     > Hi I've seen previous threads similar to this but not seen what the
>     > fix is. I am currently using Quickfix 1.12.4, when i stress the
>     > system, i.e send about 20000 trades through QF, my application
>     memory
>     > grows very rapidly. Anybody experienced the same thing?
>
>     ------------------------------------------------------------------------------
>     This SF.net email is sponsored by:
>     SourcForge Community
>     SourceForge wants to tell your story.
>     http://p.sf.net/sfu/sf-spreadtheword
>     _______________________________________________
>     Quickfix-developers mailing list
>     Qui...@li...
>     <mailto:Qui...@li...>
>     https://lists.sourceforge.net/lists/listinfo/quickfix-developers
>
>

Re: [Quickfix-developers] C++ Quickfix Memory keeps growing

[Djalma R. d. S. F. <drs...@gm...>]

Yes, this is a pseudo memory leak.

Keeping all outbound sequence numbers in maps wasn't a smart solution for
the FileStore.

To avoid development effort, you can use the other available store factories
with their pros and cons (memory, MySQL, ODBC etc).

But, what sounds to me as a solution to save the current FileStore from this
problem is the .header file in a binary format (a sequence of fixed blocks /
serialized structs). In this file QF would append and seek on demmand the
seqnum's position data (offset | message length) in corresponding .body file
(additionally a cache for the LRU would be nice, but not all data in
memory).

On Wed, Jan 21, 2009 at 5:42 PM, Alexey Zubko <
ale...@in...> wrote:

> QuickFIX Documentation:
> http://www.quickfixengine.org/quickfix/doc/html/index.html
> QuickFIX Support: http://www.quickfixengine.org/services.html
>
> That's probably the maps keeping messages for resend (store factory).
> Take a look into the /PersistMessages /configuration parameter.
> As alternative you can write your own store factory.
>
> Regards,
>        Alexey Zubko
>
>
>
>
>
> Michele Chitson wrote:
> > Hi I've seen previous threads similar to this but not seen what the
> > fix is. I am currently using Quickfix 1.12.4, when i stress the
> > system, i.e send about 20000 trades through QF, my application memory
> > grows very rapidly. Anybody experienced the same thing?
>
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by:
> SourcForge Community
> SourceForge wants to tell your story.
> http://p.sf.net/sfu/sf-spreadtheword
> _______________________________________________
> Quickfix-developers mailing list
> Qui...@li...
> https://lists.sourceforge.net/lists/listinfo/quickfix-developers
>

Re: [Quickfix-developers] C++ Quickfix Memory keeps growing

[Djalma R. d. S. F. <drs...@gm...>]

Yes, this is a pseudo memory leak.

Keeping all outbound sequence numbers in maps wasn't a smart solution for
the FileStore.

To avoid development effort, you can use the other available store factories
with their pros and cons (memory, MySQL, ODBC etc).

But, what sounds to me as a solution to save the current FileStore from this
problem is the .header file in a binary format (a sequence of fixed blocks /
serialized structs). In this file QF would append and seek on demmand the
seqnum's position data (offset | message length) in corresponding .body file
(additionally a cache for the LRU would be nice, but not all data in
memory).



On Wed, Jan 21, 2009 at 5:42 PM, Alexey Zubko <
ale...@in...> wrote:

> QuickFIX Documentation:
> http://www.quickfixengine.org/quickfix/doc/html/index.html
> QuickFIX Support: http://www.quickfixengine.org/services.html
>
> That's probably the maps keeping messages for resend (store factory).
> Take a look into the /PersistMessages /configuration parameter.
> As alternative you can write your own store factory.
>
> Regards,
>        Alexey Zubko
>
>
>
>
>
> Michele Chitson wrote:
> > Hi I've seen previous threads similar to this but not seen what the
> > fix is. I am currently using Quickfix 1.12.4, when i stress the
> > system, i.e send about 20000 trades through QF, my application memory
> > grows very rapidly. Anybody experienced the same thing?
>
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by:
> SourcForge Community
> SourceForge wants to tell your story.
> http://p.sf.net/sfu/sf-spreadtheword
> _______________________________________________
> Quickfix-developers mailing list
> Qui...@li...
> https://lists.sourceforge.net/lists/listinfo/quickfix-developers
>

Re: [Quickfix-developers] NewOrderList addgroup threw "Attempted to read or write protected memory. This is often an indication that other memory is corrupt"

[Djalma R. d. S. F. <drs...@gm...>]

This error occurs because the .NET runtime sometimes moves the objects in
memory to compact the GC Heap, this cheats the unmanaged code and leads to
an access violation. The solution is pinning the objects involved.

void addGroup( Group* group )
  { QF_STACK_TRY
    if ( !group ) { throw new System::ArgumentNullException(); }
    Group __pin *pThis = this;
    Group __pin *pGroup = group;
    pThis->unmanaged().addGroup( pGroup->unmanaged() );
    QF_STACK_CATCH
  }


On Thu, Jan 22, 2009 at 5:44 PM, Gehani, Neeraj <nee...@ci...>wrote:

> QuickFIX Documentation:
> http://www.quickfixengine.org/quickfix/doc/html/index.html
> QuickFIX Support: http://www.quickfixengine.org/services.html
>
> Here is the log of exception this threw.
>
> This is the first time this problem has surfaced. App has been used for
> almost half year. Any Ideas.
>
> 2009-01-22 13:30:06,515 [35] FATAL TradeManager - Attempted to read or
> write protected memory. This is often an indication that other memory is
> corrupt.    at FIX.FieldMap.addGroup(FieldMap* , Int32 , FieldMap* ,
> Boolean )
>   at QuickFix.Message.addGroup(Group group)
>   at FixUtil.createNewOrderList(FixRequestDetails frd)
>   at Communicator.SendNewOrderList(FixRequestDetails frd)
>   at TradeManager.Execute(Queue`1 queue)
>   at TradeManager.Manage(Object state)
>
> Code statements
>
>            int listSeqNo = 0;
>
>            QuickFix44.NewOrderList newOrderList = new
> QuickFix44.NewOrderList();
>            Hashtable listsCache = null;
>
>            QuickFix.ListID listID = null;
>
>            listsCache =
> PortfolioManager.Instance.getListsCache(frd.PortfolioID);
>
>            listID = new ListID("ABC");
>            newOrderList.setField(listID);
>
>            QuickFix.BidType bidType = new
> QuickFix.BidType(QuickFix.BidType.DISCLOSED_STYLE);
>            newOrderList.setField(bidType);
>
>            QuickFix44.NewOrderList.NoOrders sequenceGroup = null;
>
>                foreach (Order order in orders.Values)
>                {
>
>                    sequenceGroup = new
> QuickFix44.NewOrderList.NoOrders();
>                    SoupLogger.getLogger().InfoFormat("Processing
> Frd={0} Order {1}", frd.getTag66(),order.ToString());
>                    sequenceGroup.setField(new
> QuickFix.Account(order.ActMnc));
>                    sequenceGroup.setField(new
> QuickFix.ClOrdID(order.OrderId));
>                    sequenceGroup.setField(new
> QuickFix.ListSeqNo(++listSeqNo));
>                    sequenceGroup.setField(new
> QuickFix.Symbol(order.Ticker));
>                    if (order.TickerSuff != null &&
> order.TickerSuff.Length != 0 ) sequenceGroup.setField(new
> QuickFix.SymbolSfx(order.TickerSuff));
>
>                    if (order.Side == OMConstants.BUY)
>                    {
>                        sequenceGroup.setField(new
> QuickFix.Side(QuickFix.Side.BUY));
>                    }
>                    else if (order.Side == OMConstants.SELL)
>                    {
>                        sequenceGroup.setField(new
> QuickFix.Side(QuickFix.Side.SELL));
>                    }
>                    else if (order.Side == OMConstants.SELLSHORT)
>                    {
>                        sequenceGroup.setField(new
> QuickFix.Side(QuickFix.Side.SELL_SHORT));
>                    }
>                    else if (order.Side == OMConstants.SELLSHORTEXEMPT)
>                    {
>                        sequenceGroup.setField(new
> QuickFix.Side(QuickFix.Side.SELL_SHORT_EXEMPT));
>                    }
>
>                      string[] locateIdArr = order.LocateIds.Split(':');
>
> sequenceGroup.setField(FixConstants.LOCATEBROKERNAME, locateIdArr[0]);
>
> sequenceGroup.setField(FixConstants.LOCATEBROKERID, locateIdArr[1]);
>
>
>                    sequenceGroup.setField(new
> QuickFix.OrderQty(order.Quantity));
>                    sequenceGroup.setField(FixConstants.SUBLISTID,
> order.ListId);
>                    sequenceGroup.setField(FixConstants.TRADERSOEID,
> order.ClientName);
>                    sequenceGroup.setField(new
> IntField(FixConstants.SRCSYSTEMID,srcSystemId));
>                    sequenceGroup.setField(new
> IntField(FixConstants.AGGUNIT, 999);
>
>                    sequenceGroup.setField(new
> QuickFix.SenderCompID(senderCompId));
>                    sequenceGroup.setField(new
> QuickFix.TargetCompID(targetCompId));
>                    sequenceGroup.setField(new
> QuickFix.OrderID(order.OrderId));
>                    sequenceGroup.setField(new
> QuickFix.TimeInForce(TimeInForce.DAY));
>                    sequenceGroup.setField(new
> QuickFix.OrderCapacity(OrderCapacity.AGENCY));
>                    sequenceGroup.setField(new
> QuickFix.UtcTimeStampField( FixConstants.ORDERFLOWENTRYTIME,
> DateTime.UtcNow));
>
>
>
>                    newOrderList.addGroup(sequenceGroup);
>                }
>
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by:
> SourcForge Community
> SourceForge wants to tell your story.
> http://p.sf.net/sfu/sf-spreadtheword
> _______________________________________________
> Quickfix-developers mailing list
> Qui...@li...
> https://lists.sourceforge.net/lists/listinfo/quickfix-developers
>

[Quickfix-developers] NewOrderList addgroup threw "Attempted to read or write protected memory. This is often an indication that other memory is corrupt"

[Gehani, N. <nee...@ci...>]

Here is the log of exception this threw.

This is the first time this problem has surfaced. App has been used for
almost half year. Any Ideas.

2009-01-22 13:30:06,515 [35] FATAL TradeManager - Attempted to read or
write protected memory. This is often an indication that other memory is
corrupt.    at FIX.FieldMap.addGroup(FieldMap* , Int32 , FieldMap* ,
Boolean )
   at QuickFix.Message.addGroup(Group group)
   at FixUtil.createNewOrderList(FixRequestDetails frd)
   at Communicator.SendNewOrderList(FixRequestDetails frd)
   at TradeManager.Execute(Queue`1 queue)
   at TradeManager.Manage(Object state)

Code statements

            int listSeqNo = 0;

            QuickFix44.NewOrderList newOrderList = new
QuickFix44.NewOrderList();
            Hashtable listsCache = null;

            QuickFix.ListID listID = null;

            listsCache =
PortfolioManager.Instance.getListsCache(frd.PortfolioID);

            listID = new ListID("ABC");
            newOrderList.setField(listID);

            QuickFix.BidType bidType = new
QuickFix.BidType(QuickFix.BidType.DISCLOSED_STYLE);
            newOrderList.setField(bidType);

            QuickFix44.NewOrderList.NoOrders sequenceGroup = null;

                foreach (Order order in orders.Values)
                {
 
                    sequenceGroup = new
QuickFix44.NewOrderList.NoOrders();
                    SoupLogger.getLogger().InfoFormat("Processing
Frd={0} Order {1}", frd.getTag66(),order.ToString());
                    sequenceGroup.setField(new
QuickFix.Account(order.ActMnc));
                    sequenceGroup.setField(new
QuickFix.ClOrdID(order.OrderId));
                    sequenceGroup.setField(new
QuickFix.ListSeqNo(++listSeqNo));
                    sequenceGroup.setField(new
QuickFix.Symbol(order.Ticker));
                    if (order.TickerSuff != null &&
order.TickerSuff.Length != 0 ) sequenceGroup.setField(new
QuickFix.SymbolSfx(order.TickerSuff));

                    if (order.Side == OMConstants.BUY)
                    {
                        sequenceGroup.setField(new
QuickFix.Side(QuickFix.Side.BUY));
                    }
                    else if (order.Side == OMConstants.SELL)
                    {
                        sequenceGroup.setField(new
QuickFix.Side(QuickFix.Side.SELL));
                    }
                    else if (order.Side == OMConstants.SELLSHORT)
                    {
                        sequenceGroup.setField(new
QuickFix.Side(QuickFix.Side.SELL_SHORT));
                    }
                    else if (order.Side == OMConstants.SELLSHORTEXEMPT)
                    {
                        sequenceGroup.setField(new
QuickFix.Side(QuickFix.Side.SELL_SHORT_EXEMPT));
                    }

                      string[] locateIdArr = order.LocateIds.Split(':');
 
sequenceGroup.setField(FixConstants.LOCATEBROKERNAME, locateIdArr[0]);
 
sequenceGroup.setField(FixConstants.LOCATEBROKERID, locateIdArr[1]);


                    sequenceGroup.setField(new
QuickFix.OrderQty(order.Quantity));
                    sequenceGroup.setField(FixConstants.SUBLISTID,
order.ListId);
                    sequenceGroup.setField(FixConstants.TRADERSOEID,
order.ClientName);
                    sequenceGroup.setField(new
IntField(FixConstants.SRCSYSTEMID,srcSystemId));
                    sequenceGroup.setField(new
IntField(FixConstants.AGGUNIT, 999);

                    sequenceGroup.setField(new
QuickFix.SenderCompID(senderCompId));
                    sequenceGroup.setField(new
QuickFix.TargetCompID(targetCompId));
                    sequenceGroup.setField(new
QuickFix.OrderID(order.OrderId));
                    sequenceGroup.setField(new
QuickFix.TimeInForce(TimeInForce.DAY));
                    sequenceGroup.setField(new
QuickFix.OrderCapacity(OrderCapacity.AGENCY));
                    sequenceGroup.setField(new
QuickFix.UtcTimeStampField( FixConstants.ORDERFLOWENTRYTIME,
DateTime.UtcNow));



                    newOrderList.addGroup(sequenceGroup);
                }

Re: [Quickfix-developers] C++ Quickfix Memory keeps growing

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

That's probably the maps keeping messages for resend (store factory).
Take a look into the /PersistMessages /configuration parameter.
As alternative you can write your own store factory.

Regards,
	Alexey Zubko





Michele Chitson wrote:
> Hi I've seen previous threads similar to this but not seen what the 
> fix is. I am currently using Quickfix 1.12.4, when i stress the 
> system, i.e send about 20000 trades through QF, my application memory 
> grows very rapidly. Anybody experienced the same thing? 

[Quickfix-developers] C++ Quickfix Memory keeps growing

[Michele C. <mch...@gm...>]

Hi I've seen previous threads similar to this but not seen what the fix is.
I am currently using Quickfix 1.12.4, when i stress the system, i.e send
about 20000 trades through QF, my application memory grows very rapidly.
Anybody experienced the same thing?

Re: [Quickfix-developers] Unable to connect to Server using QuickFixfor .Net

[Romman, G. <GR...@ep...>]

Hi Nikhil,

Is it possible the server requires logon credentials such as userid and
password?  The documention FAQ provides a sample of how to add fields to
a logon message.  Here also is a working snippet:

class YourApplication : Application
 public void toAdmin(QuickFix.Message message, SessionID sessionID)
 {
     // Add custom fields if this is a logon message
      QuickFix.Message.Header hdr = message.getHeader();
      if (hdr.getField(MsgType.FIELD).Equals(MsgType.LOGON))
      {
      message.setField(new Username("userid"));
      message.setField(new Password("password"));
      }
 }

Regards,
Greg

-----Original Message-----
From: Nikhil Chitnis [mailto:nik...@gm...] 
Sent: Wednesday, January 21, 2009 4:20 AM
To: qui...@li...
Subject: [Quickfix-developers] Unable to connect to Server using
QuickFixfor .Net

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



This email contains confidential information that is proprietary to EPCOR and its subsidiary companies in all respects.  This information is intended only for the person(s) named in the destination address.  Unauthorized distribution, copying or disclosure is strictly prohibited.  If you receive this e-mail in error please delete it immediately.

[Quickfix-developers] Unable to connect to Server using QuickFix for .Net

[Nikhil C. <nik...@gm...>]

Hi,
I am unable to connect to the server using QuickFix for .Net. Following is
the code I am using -

static void Main(string[] args)
{
try
{
SessionSettings settings = new SessionSettings(@"C:\TestFIX\Test.cfg");
Application application = new Application();
FileStoreFactory storeFactory = new FileStoreFactory( settings );
ScreenLogFactory logFactory = new ScreenLogFactory( settings );
MessageFactory messageFactory = new DefaultMessageFactory();
SocketInitiator testcos = new SocketInitiator(application, storeFactory,
settings, logFactory, messageFactory);

testcos.start();
bool bl = testcos.isLoggedOn();
Console.Clear();
Console.WriteLine("press <enter> to quit");
Console.Read();
testcos.stop();
}
catch (ConfigError e)
{
Console.WriteLine(e);
}
}

When I run the application, I am getting following on the console -

<20090121-10:38:52, FIX.4.2:LTINTTEST->ONYX, outgoing>
  (8=FIX.4.2☺9=69☺35=A☺34=10☺49=LTINTTEST☺52=20090121-10:38:52.222☺56=ONYX☺98=0☺108=30☺10=068☺)
<20090121-10:38:52, FIX.4.2:LTINTTEST->ONYX, event>
  (Initiated logon request)
<20090121-10:38:52, FIX.4.2:LTINTTEST->ONYX, event>
  (Socket Error: Connection reset by peer.)
<20090121-10:38:52, FIX.4.2:LTINTTEST->ONYX, event>
  (Disconnecting)
<20090121-10:38:52, FIX.4.2:LTINTTEST->ONYX, event>
  (Connecting to 66.150.110.105 on port 4759)
<20090121-10:38:53, FIX.4.2:LTINTTEST->ONYX, outgoing>
  (8=FIX.4.2☺9=69☺35=A☺34=11☺49=LTINTTEST☺52=20090121-10:38:53.222☺56=ONYX☺98=0☺108=30☺10=070☺)
<20090121-10:38:53, FIX.4.2:LTINTTEST->ONYX, event>
  (Initiated logon request)
<20090121-10:38:56, FIX.4.2:LTINTTEST->ONYX, event>
  (Socket Error: Connection reset by peer.)
<20090121-10:38:56, FIX.4.2:LTINTTEST->ONYX, event>
  (Disconnecting)

Can anyone help me in this?

Thanks in advance.

Nikhil

[Quickfix-developers] bug tracker

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

Is quickfix bug tracker down?
Where I can open issues?


-- 


Regards,
	Alexey Zubko

Re: [Quickfix-developers] Acceptor broadcasting?

[V. K. <vk...@el...>]

Thanks for the replies.

Creating multiple "spinning/waiting" threads doesn't sound very appealing.

Microsoft have a neat solution with their Parallel Extensions Framework 
(PFX) that's currently in CTP (a part of .NET 4.0 according to the VS2010 
CTP).

Parallel.ForEach(acceptor.getSessions(), session =>
{
    Session.SendToTarget(message,session);
}

I haven't measured the speed increase (if any) but I hope MS make this idea 
work for us .NET programmers.

nisbus
 

[Quickfix-developers] Building in VS8 .NET

[V. K. <vk...@el...>]

Hi, 

I just tried building quickfix using VS8.

1. downloaded the source
2. edited the quickfix\src\config_windows.h file to include "#define 
HAVE_MSSQL 1" (per instructions from the documentation 
http://www.quickfixengine.org/quickfix/doc/html/building.html).
3. opened quickfix_v8.sln from the quickfix directory. (was asked to 
convert the project since it was created with an earlier version of VS).
4. built the solution successfully.

If I run the sample executor I get the following error:

System.BadImageFormatException was unhandled
  Message="Could not load file or assembly 'quickfix_net, 
Version=1.0.3303.42637, Culture=neutral, PublicKeyToken=null' or one of its 
dependencies. An attempt was made to load a program with an incorrect 
format."
  Source="executor_csharp"
  FileName="quickfix_net, Version=1.0.3303.42637, Culture=neutral, 
PublicKeyToken=null"
  FusionLog="=== Pre-bind state information ===\r\nLOG: User = ****\r\nLOG: 
DisplayName = quickfix_net, Version=1.0.3303.42637, Culture=neutral, 
PublicKeyToken=null\n (Fully-specified)\r\nLOG: Appbase = 
file:///C:/Users/***/Documents/quickfix/bin/debug/executor_csharp/\r\nLOG: 
Initial PrivatePath = NULL\r\nCalling assembly : executor_csharp, 
Version=1.0.3303.43050, Culture=neutral, 
PublicKeyToken=null.\r\n===\r\nLOG: This bind starts in default load 
context.\r\nLOG: No application configuration file found.\r\nLOG: Using 
machine configuration file from 
C:\\Windows\\Microsoft.NET\\Framework64\\v2.0.50727\\config\\machine.config.
\r\nLOG: Policy not being applied to reference at this time (private, 
custom, partial, or location-based assembly bind).\r\nLOG: Attempting 
download of new URL 
file:///C:/Users/***/Documents/quickfix/bin/debug/executor_csharp/quickfix_n
et.DLL.\r\nERR: Failed to complete setup of assembly (hr = 0x8007000b). 
Probing terminated.\r\n"
  StackTrace:
       at executor_csharp.Executor.Main(String[] args)
       at System.AppDomain._nExecuteAssembly(Assembly assembly, String[] 
args)
       at 
Microsoft.VisualStudio.HostingProcess.HostProc.RunUsersAssembly()
       at System.Threading.ExecutionContext.Run(ExecutionContext 
executionContext, ContextCallback callback, Object state)
       at System.Threading.ThreadHelper.ThreadStart()
  InnerException: 

Stack trace:

   at executor_csharp.Executor.Main(String[] args)
   at System.AppDomain._nExecuteAssembly(Assembly assembly, String[] args)
   at Microsoft.VisualStudio.HostingProcess.HostProc.RunUsersAssembly()
   at System.Threading.ExecutionContext.Run(ExecutionContext 
executionContext, ContextCallback callback, Object state)
   at System.Threading.ThreadHelper.ThreadStart()

I also tried creating a new console project that referenced 
quickfix\src\.NET\debug\vs8\quickfix_net.dll but the same error appeared.

I'm running on Vista 64bit and was wondering if I need to make so special 
consessions for that?
I tried building the executor as a 64bit app and changed it to .NET v3.5 
but the same error.

Anyone else tried this?

Thanks,
  nisbus

 

[Quickfix-developers] Message Buffer Size

[Vincent P. <vpr...@ph...>]

When my test program sends 50,000 FIX trades, some of the execution  
report messages get lost on the return trip.  This doesn't happen  
when the test size is smaller, like 20,000 trades.  Some ERs appear  
to be lost and my theory is that the quickfix buffer is overflowing.   
I am using the file store.    How can I increase the size of the  
buffer, or guarantee that messages won't be dropped?

I am setting PersistMessages in my .cfg file of both the initiator  
and the acceptor ( attached ).  Still, I am losing messages.  Is  
there something else I need to do?

-- 
Vincent

[DEFAULT]
ConnectionType=initiator
HeartBtInt=30
ReconnectInterval=1
FileStorePath=sellerstore
FileLogPath=log
StartTime=00:00:00
EndTime=00:00:00
UseDataDictionary=N
SocketConnectHost=localhost
PersistMessages=Y


[DEFAULT]
ConnectionType=acceptor
SocketAcceptPort=8003
SocketReuseAddress=Y
StartTime=00:00:00
EndTime=00:00:00
PersistMessages=Y

[Quickfix-developers] Acceptor broadcasting?

[V. K. <vk...@el...>]

Hi,

I just started using the .NET quickfix lib and finding it surprisingly 
simple to use.

I have an acceptor that my clients connect to.
The acceptor is getting data from an exchange and then needs to forward the 
messages to all logged on clients (the exchange isn't using FIX).

When I convert messages from the exchange to FIX and the acceptor picks it 
up it seems to me that I would have to loop through the 
sessions and send the message to each client. This seems like an 
unnecessarily arduous task so I wondering if there is a way to have the 
acceptor broadcast messages to all connected clients?

this is how I'm doing this now:

ArrayList sessions = acceptor.GetSessions();
for (int i=0; i< sessions.Count;i++)
{
    Session.sendMessageToTargets(msg,((SessionID)sessions[i]));
}

I would like to be able to just call acceptor.SendAll(msg); instead.

Thanks,
  nisbus

 

Re: [Quickfix-developers] How do I change the field of a message from the .net api?

[Andrei G. <an...@gm...>]

On Fri, Jan 9, 2009 at 1:54 AM, Wilhelm Thomas <th...@cu...> wrote:
> QuickFIX Documentation: http://www.quickfixengine.org/quickfix/doc/html/index.html
> QuickFIX Support: http://www.quickfixengine.org/services.html
>
>
> hello
>
> Happy new year to all!
>
> I'm using the .net api
>
> I'm working on a gateway, I receive a message from my acceptor function
>
> fromApp(
>
> Message message, SessionID sessionID)
>
> 1) I basically need to change some fields from this message and forward back the message to my Initiator
>
> I'm basically doing this:
>
>
> message.removeField((int)49);
> message.removeField(56);
> Console.Out.WriteLine(">>>> m=" + message.ToString());
> message.setField(49, "MAXP2"); // SenderCompID
> message.setField(56, "MAX2PRICE"); // TargetCompID
> Console.Out.WriteLine(">>>> m=" + message.ToString());
> message.setString(49, "MAXP2"); // SenderCompID
> message.setString(56, "MAX2PRICE"); // TargetCompID
> Console.Out.WriteLine(">>>> m=" + message.ToString());
>
>
> but when I call some ,remove, replace, setField, setString, nothing change the > fields 49 and 56?

You are setting header fields into the body. This is a no-no. The
correct way to do it would be:

message.getHeader().setString(49, "MAXP2"); // SenderCompID
message.getHeader().setString(56, "MAX2PRICE"); // TargetCompID

Anyway, keep in mind that if you want to send this message via
QuickFIX (i.e. SendToTarget), MAXP2-MAX2PRICE must be a valid session
defined in your coniguration file.

>
> 2) I also try to loop over the fields and create a new message but I do not get the correct field
>
> System.Collections.
>
> IEnumerator enum_ = message.GetEnumerator();
> QuickFix.Field key = null;
> while(enum_.MoveNext()){
>   key = (Field)enum_.Current;
>   message2.setField(key.getField(), (string)key.getObject());
> }
>
> Here is what I see once this while execute:
>
> +  message {8=FIX.4.2 9=225 35=V 34=2 49=BLABLA 52=20090109-03:49:01.464 56=GATEWAY146=1 55=6E Sep08 48=6E012008090000003232 167=FUT 200=200809 202=0 207=TTSIM 262=1/8/2009 7:49:01 PM172 263=1 264=10 265=1 266=Y 267=3 269=0 269=1 269=2 10=192 } QuickFix.Message {QuickFix42.MarketDataRequest}
>
> +  message2 {8=FIX.4.2 9=69 35=V 146=1 262=1/8/2009 7:49:01 PM172 263=1 264=10 265=1 266=Y 267=3 10=251 } QuickFix42.MarketDataRequest
>
> Which is weird as message2 isn't at all like message?
>
> Any idea why?

1) You are not iterating over header/trailer fields.
2) You won't get fields in repeating group, as they are kept in
separate structures.

Re: [Quickfix-developers] How do I change the field of a message from the .net api?

[Peter P. <pet...@ho...>]

Dear Wilhem,

I'm new to QuickFix but maybe I can help? As someone coming new to it, I also found that Quickfix was very tricky to use; it displays truly bizarre behaviour at times!

I do not believe that you can build a message manually in this way, because Quickfix wants to understand the structure of what you are building (with repeating groups etc) and will probably just throw out things that it does not like.

When I needed to do what you are trying, I decided to build a new message by knowing what fields should be present in the incoming message and getting those values explicitly then rebuilding the groups using the QF-standard way, not by simply setting fields manually.

Yours,

Dan

Date: Thu, 8 Jan 2009 19:54:41 -0800
From: th...@cu...
To: qui...@li...
Subject: [Quickfix-developers] How do I change the field of a message from	the .net api?


hello
 
Happy new year to all!
 
I'm using the .net api
 
I'm working on a gateway, I receive a message from my acceptor function 
fromApp(Message message, SessionID sessionID) 
1) I basically need to change some fields from this message and forward back the message to my Initiator
I'm basically doing this:
 
message.removeField((int)49);
message.removeField(56);
Console.Out.WriteLine(">>>> m=" + message.ToString());

message.setField(49, "MAXP2"); // SenderCompID

message.setField(56, "MAX2PRICE"); // TargetCompID

Console.Out.WriteLine(">>>> m=" + message.ToString());

message.setString(49, "MAXP2"); // SenderCompID

message.setString(56, "MAX2PRICE"); // TargetCompID

Console.Out.WriteLine(">>>> m=" + message.ToString());

 
but when I call some ,remove, replace, setField, setString, nothing change the fields 49 and 56?
2) I also try to loop over the fields and create a new message but I do not get the correct field

System.Collections.IEnumerator enum_ = message.GetEnumerator();
QuickFix.Field key = null;

while(enum_.MoveNext()){
  key = (Field)enum_.Current;

  message2.setField(key.getField(), (string)key.getObject());
}
Here is what I see once this while execute:
+  message {8=FIX.4.2
9=225
35=V
34=2
49=BLABLA
52=20090109-03:49:01.464
56=GATEWAY146=1
55=6E Sep08
48=6E012008090000003232
167=FUT
200=200809
202=0
207=TTSIM
262=1/8/2009 7:49:01 PM172
263=1
264=10
265=1
266=Y
267=3
269=0
269=1
269=2
10=192
} QuickFix.Message {QuickFix42.MarketDataRequest}


+  message2 {8=FIX.4.2
9=69
35=V
146=1
262=1/8/2009 7:49:01 PM172
263=1
264=10
265=1
266=Y
267=3
10=251
} QuickFix42.MarketDataRequest

Which is weird as message2 isn't at all like message? 
Any idea why?
thanks a lot for your help
Wilhem
_________________________________________________________________
Choose the perfect PC or mobile phone for you
http://clk.atdmt.com/UKM/go/130777504/direct/01/

[Quickfix-developers] Does anyone else find that Quickfix connection management is not robust?

[Peter P. <pet...@ho...>]

I have three applications that want to talk to each other in a chain. Because they are three separate applications (not three DLLs in the same process), all three instances of Quickfix should be completely independent.

So, I have an initiator A that connects to a hybrid (acceptor+initiator) application B, that connects to an acceptor-only application C

If I start them all in quick succession on the same Windows machine (e.g. by running a batch file), B successfully connects to C, but A never manages to connect to B. I get the admin notification in A saying that it sent the Logon message, but nothing in B to say that it received it.

If I start C, then pause for a few seconds, then start B, then pause for a few seconds, then start A, the logons all work properly.

If I enable file logging, to see what events are being posted. Absolutely *no* events get posted when I do the batch version but the slow-and-steady version posts events and logs on.

Has anyone else seen behaviour like this? Is Quickfix known to be a bit sensitive (flaky even?) when starting up?

Thanks,

Dan


_________________________________________________________________
Imagine a life without walls.  See the possibilities
http://clk.atdmt.com/UKM/go/122465943/direct/01/

[Quickfix-developers] How do I change the field of a message from the .net api?

[Wilhelm T. <th...@cu...>]

 hello

Happy new year to all!

I'm using the .net api

I'm working on a gateway, I receive a message from my acceptor function

fromApp(Message message, SessionID sessionID)

1) I basically need to change some fields from this message and forward back
the message to my Initiator

I'm basically doing this:

message.removeField((int)49);
message.removeField(56);
Console.Out.WriteLine(">>>> m=" + message.ToString());
message.setField(49, "MAXP2"); // SenderCompID
message.setField(56, "MAX2PRICE"); // TargetCompID
Console.Out.WriteLine(">>>> m=" + message.ToString());
message.setString(49, "MAXP2"); // SenderCompID
message.setString(56, "MAX2PRICE"); // TargetCompID
Console.Out.WriteLine(">>>> m=" + message.ToString());


but when I call some ,remove, replace, setField, setString, nothing change
the fields 49 and 56?
2) I also try to loop over the fields and create a new message but I do not
get the correct field

System.Collections.IEnumerator enum_ = message.GetEnumerator();
QuickFix.Field key = null;
while(enum_.MoveNext()){
  key = (Field)enum_.Current;
  message2.setField(key.getField(), (string)key.getObject());
}

Here is what I see once this while execute:

+  message {8=FIX.4.2
9=225
35=V
34=2
49=BLABLA
52=20090109-03:49:01.464
56=GATEWAY146=1
55=6E
Sep08
48=6E012008090000003232
167=FUT
200=200809
202=0
207=TTSIM
262=1/8/2009
7:49:01
PM172
263=1
264=10
265=1
266=Y
267=3
269=0
269=1
269=2
10=192
}
QuickFix.Message
{QuickFix42.MarketDataRequest}

+  message2 {8=FIX.4.2
9=69
35=V
146=1
262=1/8/2009 7:49:01
PM172
263=1
264=10
265=1
266=Y
267=3
10=251
} QuickFix42.MarketDataRequest

Which is weird as message2 isn't at all like message?

Any idea why?

thanks a lot for your help

Wilhem

[Quickfix-developers] Could any one help me? Can't start QF in java

[Edward_Cao <Cha...@gl...>]

I've compiled the DLL in vc but got a crash in java when using jni

Error Log file:

#
# An unexpected error has been detected by Java Runtime Environment:
#
#  EXCEPTION_ACCESS_VIOLATION (0xc0000005) at pc=0x100b0926, pid=2848,
tid=3824
#
# Java VM: Java HotSpot(TM) Client VM (11.0-b16 mixed mode windows-x86)
# Problematic frame:
# C  [quickfix_jni.dll+0xb0926]
#
# If you would like to submit a bug report, please visit:
#   http://java.sun.com/webapps/bugreport/crash.jsp
# The crash happened outside the Java Virtual Machine in native code.
# See problematic frame for where to report the bug.
#

---------------  T H R E A D  ---------------

Current thread (0x003b6c00):  JavaThread "main" [_thread_in_native, id=3824,
stack(0x008c0000,0x00910000)]

siginfo: ExceptionCode=0xc0000005, reading address 0x00000008

Registers:
EAX=0x00000004, EBX=0x06b95648, ECX=0x00000004, EDX=0x0090f3f8
ESP=0x0090f23c, EBP=0x0090f314, ESI=0x0090f3e8, EDI=0x0090f314
EIP=0x100b0926, EFLAGS=0x00010216

Top of Stack: (sp=0x0090f23c)
0x0090f23c:   0090f3cc 0090f3e8 06b95648 cccccccc
0x0090f24c:   cccccccc cccccccc cccccccc cccccccc
0x0090f25c:   cccccccc cccccccc cccccccc cccccccc
0x0090f26c:   cccccccc cccccccc cccccccc cccccccc
0x0090f27c:   cccccccc cccccccc cccccccc cccccccc
0x0090f28c:   cccccccc cccccccc cccccccc cccccccc
0x0090f29c:   cccccccc cccccccc cccccccc cccccccc
0x0090f2ac:   cccccccc cccccccc cccccccc cccccccc 

Instructions: (pc=0x100b0926)
0x100b0916:   00 00 b8 cc cc cc cc f3 ab 59 89 4d f8 8b 45 f8
0x100b0926:   8b 48 04 51 e8 e4 60 fd ff 83 c4 04 5f 5e 5b 81 


Stack: [0x008c0000,0x00910000],  sp=0x0090f23c,  free space=316k
Native frames: (J=compiled Java code, j=interpreted, Vv=VM code, C=native
code)
C  [quickfix_jni.dll+0xb0926]
C  [quickfix_jni.dll+0x116986]
C  [quickfix_jni.dll+0x115f7a]
C  [quickfix_jni.dll+0x115314]
C  [quickfix_jni.dll+0x114526]
C  [quickfix_jni.dll+0x162cfd]
C  [quickfix_jni.dll+0x13f9f6]
C  [quickfix_jni.dll+0x13f78e]
C  [quickfix_jni.dll+0xd3c1e]
j  quickfix.SessionSettings.create(Ljava/io/InputStream;)V+0
j  quickfix.SessionSettings.<init>(Ljava/io/InputStream;)V+6
j  MyClass.main([Ljava/lang/String;)V+31
v  ~StubRoutines::call_stub

Java frames: (J=compiled Java code, j=interpreted, Vv=VM code)
j  quickfix.SessionSettings.create(Ljava/io/InputStream;)V+0
j  quickfix.SessionSettings.<init>(Ljava/io/InputStream;)V+6
j  MyClass.main([Ljava/lang/String;)V+31
v  ~StubRoutines::call_stub

---------------  P R O C E S S  ---------------

Java Threads: ( => current thread )
  0x0aadc000 JavaThread "Low Memory Detector" daemon [_thread_blocked,
id=1308, stack(0x0ae90000,0x0aee0000)]
  0x0aad9c00 JavaThread "CompilerThread0" daemon [_thread_blocked, id=1724,
stack(0x0ae40000,0x0ae90000)]
  0x0aacc400 JavaThread "JDWP Command Reader" daemon [_thread_in_native,
id=932, stack(0x0adf0000,0x0ae40000)]
  0x0aacac00 JavaThread "JDWP Event Helper Thread" daemon [_thread_blocked,
id=2780, stack(0x0ada0000,0x0adf0000)]
  0x0aac8400 JavaThread "JDWP Transport Listener: dt_socket" daemon
[_thread_blocked, id=380, stack(0x0ad50000,0x0ada0000)]
  0x0aabd400 JavaThread "Attach Listener" daemon [_thread_blocked, id=1964,
stack(0x0ac70000,0x0acc0000)]
  0x0aabc000 JavaThread "Signal Dispatcher" daemon [_thread_blocked,
id=2464, stack(0x0ac20000,0x0ac70000)]
  0x0aaac400 JavaThread "Finalizer" daemon [_thread_blocked, id=3776,
stack(0x0abd0000,0x0ac20000)]
  0x0aaa7800 JavaThread "Reference Handler" daemon [_thread_blocked,
id=2104, stack(0x0ab80000,0x0abd0000)]
=>0x003b6c00 JavaThread "main" [_thread_in_native, id=3824,
stack(0x008c0000,0x00910000)]

Other Threads:
  0x0aaa2c00 VMThread [stack: 0x0ab30000,0x0ab80000] [id=2704]
  0x0aadd800 WatcherThread [stack: 0x0aee0000,0x0af30000] [id=2068]

VM state:not at safepoint (normal execution)

VM Mutex/Monitor currently owned by a thread: None

Heap
 def new generation   total 960K, used 362K [0x02990000, 0x02a90000,
0x02e70000)
  eden space 896K,  40% used [0x02990000, 0x029eaa40, 0x02a70000)
  from space 64K,   0% used [0x02a70000, 0x02a70000, 0x02a80000)
  to   space 64K,   0% used [0x02a80000, 0x02a80000, 0x02a90000)
 tenured generation   total 4096K, used 0K [0x02e70000, 0x03270000,
0x06990000)
   the space 4096K,   0% used [0x02e70000, 0x02e70000, 0x02e70200,
0x03270000)
 compacting perm gen  total 12288K, used 2071K [0x06990000, 0x07590000,
0x0a990000)
   the space 12288K,  16% used [0x06990000, 0x06b95d70, 0x06b95e00,
0x07590000)
No shared spaces configured.

Dynamic libraries:
0x00400000 - 0x00424000 	C:\Program Files\Java\jre6\bin\javaw.exe
0x7c900000 - 0x7c9b0000 	C:\WINDOWS\system32\ntdll.dll
0x7c800000 - 0x7c8f4000 	C:\WINDOWS\system32\kernel32.dll
0x77dd0000 - 0x77e6b000 	C:\WINDOWS\system32\ADVAPI32.dll
0x77e70000 - 0x77f01000 	C:\WINDOWS\system32\RPCRT4.dll
0x77d40000 - 0x77dd0000 	C:\WINDOWS\system32\USER32.dll
0x77f10000 - 0x77f58000 	C:\WINDOWS\system32\GDI32.dll
0x76390000 - 0x763ad000 	C:\WINDOWS\system32\IMM32.DLL
0x629c0000 - 0x629c9000 	C:\WINDOWS\system32\LPK.DLL
0x74d90000 - 0x74dfb000 	C:\WINDOWS\system32\USP10.dll
0x77c10000 - 0x77c68000 	C:\WINDOWS\system32\msvcrt.dll
0x7c340000 - 0x7c396000 	C:\Program Files\Java\jre6\bin\msvcr71.dll
0x6d800000 - 0x6da56000 	C:\Program Files\Java\jre6\bin\client\jvm.dll
0x76b40000 - 0x76b6d000 	C:\WINDOWS\system32\WINMM.dll
0x6d280000 - 0x6d288000 	C:\Program Files\Java\jre6\bin\hpi.dll
0x76bf0000 - 0x76bfb000 	C:\WINDOWS\system32\PSAPI.DLL
0x6d370000 - 0x6d399000 	C:\Program Files\Java\jre6\bin\jdwp.dll
0x6d6a0000 - 0x6d6a6000 	C:\Program Files\Java\jre6\bin\npt.dll
0x6d7b0000 - 0x6d7bc000 	C:\Program Files\Java\jre6\bin\verify.dll
0x6d320000 - 0x6d33f000 	C:\Program Files\Java\jre6\bin\java.dll
0x6d7f0000 - 0x6d7ff000 	C:\Program Files\Java\jre6\bin\zip.dll
0x6d1f0000 - 0x6d1f7000 	C:\Program Files\Java\jre6\bin\dt_socket.dll
0x71ab0000 - 0x71ac7000 	C:\WINDOWS\system32\WS2_32.dll
0x71aa0000 - 0x71aa8000 	C:\WINDOWS\system32\WS2HELP.dll
0x71a50000 - 0x71a8f000 	C:\WINDOWS\System32\mswsock.dll
0x76f20000 - 0x76f47000 	C:\WINDOWS\system32\DNSAPI.dll
0x76fb0000 - 0x76fb8000 	C:\WINDOWS\System32\winrnr.dll
0x76f60000 - 0x76f8c000 	C:\WINDOWS\system32\WLDAP32.dll
0x76fc0000 - 0x76fc6000 	C:\WINDOWS\system32\rasadhlp.dll
0x662b0000 - 0x66308000 	C:\WINDOWS\system32\hnetcfg.dll
0x71a90000 - 0x71a98000 	C:\WINDOWS\System32\wshtcpip.dll
0x10000000 - 0x101d2000 	C:\quickfix\lib\debug\quickfix_jni.dll
0x774e0000 - 0x7761c000 	C:\WINDOWS\system32\ole32.dll
0x77120000 - 0x771ac000 	C:\WINDOWS\system32\OLEAUT32.dll
0x10480000 - 0x1053c000 	C:\WINDOWS\system32\MSVCP71D.dll
0x10200000 - 0x10287000 	C:\WINDOWS\system32\MSVCR71D.dll

VM Arguments:
jvm_args:
-agentlib:jdwp=transport=dt_socket,suspend=y,address=localhost:3533
-Dfile.encoding=Cp1252 
java_command: MyClass conf.ini
Launcher Type: SUN_STANDARD

Environment Variables:
JAVA_HOME=C:\Java\jdk1.6.0_11
CLASSPATH=.;C:\Java\jdk1.6.0_11\lib;C:\Java\jdk1.6.0_11\lib\debug
PATH=C:/Program Files/Java/jre6/bin/client;C:/Program
Files/Java/jre6/bin;C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\Program
Files\ATI Technologies\ATI Control Panel;C:\GL Tools\apitools\bin;C:\GL
Tools\apitools\tools;C:\apache-ant-1.7.1\bin;C:\Java\jdk1.6.0_11\bin;c:\quickfix\lib;c:\quickfix\lib\debug
USERNAME=ccao
OS=Windows_NT
PROCESSOR_IDENTIFIER=x86 Family 15 Model 4 Stepping 3, GenuineIntel



---------------  S Y S T E M  ---------------

OS: Windows XP Build 2600 Service Pack 2

CPU:total 2 (1 cores per cpu, 2 threads per core) family 15 model 4 stepping
3, cmov, cx8, fxsr, mmx, sse, sse2, sse3, ht

Memory: 4k page, physical 1046620k(232180k free), swap 2517564k(1845796k
free)

vm_info: Java HotSpot(TM) Client VM (11.0-b16) for windows-x86 JRE
(1.6.0_11-b03), built on Nov 10 2008 02:15:12 by "java_re" with MS VC++ 7.1

time: Thu Jan 08 17:38:35 2009
elapsed time: 0 seconds


-- 
View this message in context: http://www.nabble.com/Could-any-one-help-me--Can%27t-start-QF-in-java-tp21347178p21347178.html
Sent from the QuickFIX - Dev mailing list archive at Nabble.com.

Re: [Quickfix-developers] FIX 5.0 and FAST support?

[Antonio C. <ant...@ga...>]

Hello, Oren.

First of all let me wish A Happy New Year to you and all the people in 
this newsgroup.

In your email message below dated Jan. 2008 you proposed a possible date 
of 2nd half 2008 for QuickFIX supporting FIX 5.0.

So, are we close to a release date ?

Best regards,

AntonioCaroselli
www.gatelab.com



Oren Miller wrote:
> QuickFIX Documentation: 
> http://www.quickfixengine.org/quickfix/doc/html/index.html
> QuickFIX Support: http://www.quickfixengine.org/services.html
>
> FIX 5.0 should be available by then.  I can't promise the same for  
> FAST however.
>
> --oren
>
> On Jan 10, 2008, at 12:34 PM, Scott Mitchell wrote:
>
>> QuickFIX Documentation: http://www.quickfixengine.org/quickfix/doc/ 
>> html/index.html
>> QuickFIX Support: http://www.quickfixengine.org/services.html
>>
>> Greetings,
>>
>> Is there any update available on QuickFIX support for FIX 5.0 and/or
>> FAST?
>>
>> We are going to need both for various projects in the second half of
>> this year and I need to decide if we will be able to stick with  
>> QuickFIX
>> (which we're currently extremely happy with) or if I need to start
>> investigating alternatives.
>>
>> Cheers,
>>
>>     Scott
>>
>> ---------------------------------------------------------------------- 
>> ---
>> Check out the new SourceForge.net Marketplace.
>> It's the best place to buy or sell services for
>> just about anything Open Source.
>> http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/ 
>> marketplace
>> _______________________________________________
>> Quickfix-developers mailing list
>> Qui...@li...
>> https://lists.sourceforge.net/lists/listinfo/quickfix-developers
>>
>
>
> -------------------------------------------------------------------------
> Check out the new SourceForge.net Marketplace.
> It's the best place to buy or sell services for
> just about anything Open Source.
> http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace 
>
> _______________________________________________
> Quickfix-developers mailing list
> Qui...@li...
> https://lists.sourceforge.net/lists/listinfo/quickfix-developers
>

[Quickfix-developers] ThreadedSocketInitiator::socketThread exception 1.12.4 revision 1906

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

There is a bug in version 1.12.4 revision 1906:
pInitiator->getLog() could return NULL (initiator without log factory).

I got exception at least in one place, may be there are other files:

THREAD_PROC ThreadedSocketInitiator::socketThread( void* p )
{ QF_STACK_TRY

.....

  if( !pConnection->connect() )
  {
// exception is here:
    pInitiator->getLog()->onEvent( "Connection failed" );



-- 


Regards,
	Alexey Zubko


Head Of The Core Development
Infinium Capital Corp, 106 Front Street East, Suite 200, Toronto, ON M5A 1E1
Phone: (416) 360-5228

Re: [Quickfix-developers] Moving QuickFix to production computer from development computer

[Francis G. <fr...@at...>]

Douglas,

Did you install the C++ libraries on your server? 

vcredist_x86 (SP1).exe

Francis

[Quickfix-developers] ConfigError Duplicate Session...

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

Oren, Quickfix developers,

I used to modify the settings parameters before running 
initiator/acceptor, but the new feature (revision 1966) throws 
ConfigError exception breaks that completely.

That could be nice to have such a feature to modify parameters. Is there 
are chance to get a non const Dictionary& get() method or something?

For example, I append the current date to the file paths, so the files 
are put into the write directory:

        m_settings = new FIX::SessionSettings(config_file);

        time_t        ltime = time(0);
        struct tm    *newtime = gmtime(&ltime);
        const int DTSTRSIZE     = 50;
        char    timestr[DTSTRSIZE+1];
        strftime(timestr, DTSTRSIZE, "\\%Y%m%d", newtime);

        typedef std::set<FIX::SessionID>    TSessionsSet;
        TSessionsSet            TheSet = m_settings->getSessions();
        for(TSessionsSet::iterator    it = TheSet.begin(); it != 
TheSet.end(); it++)
        {
            FIX::SessionID        SessionID = *it;
// that could be nice to get reference to the Dictionary instead here:
            FIX::Dictionary        Dictionary = m_settings->get(SessionID);

            std::string path = Dictionary.getString(FIX::FILE_LOG_PATH);
            FIX::file_mkdir(path.c_str());
            path += timestr;
            if(suffix)
            {
                path += suffix;
            }
            Dictionary.setString(FIX::FILE_LOG_PATH, path);

            path = Dictionary.getString(FIX::FILE_STORE_PATH);
            FIX::file_mkdir(path.c_str());
            path += timestr;
            if(suffix)
            {
                path += suffix;
            }
            Dictionary.setString(FIX::FILE_STORE_PATH, path);

// here is I got the exception :(
            m_settings->set(SessionID, Dictionary);
        }
    }
        m_factory = new FIX::QFMyFileStoreFactory(*m_settings);
        m_logFactory = new FIX::FileLogFactory(*m_settings);
        m_initiator = new FIX::ThreadedSocketInitiator(*this, 
*m_factory, *m_settings, *m_logFactory);
......


Thank you in advance.



-- 


Regards,
	Alexey Zubko

Re: [Quickfix-developers] WaitForSingleObject deadlock

[Malinka R. <ael...@gm...>]

Thank you for the response, and you're exactly right that is what was
happening however I was able to fix it by moving my cleanup to a
slightly different situation.  You're right it is very frustrating
that windows provides a DllMain call which could be very useful for
Initialization and Cleanup but then cripple it with things like the
threading restrictions, and not being able to call into other (even
windows) libraries.

On Tue, Jan 6, 2009 at 09:14, Peter Piper <pet...@ho...> wrote:
> QuickFIX Documentation: http://www.quickfixengine.org/quickfix/doc/html/index.html
> QuickFIX Support: http://www.quickfixengine.org/services.html
>
>
> I don't know about the internals of QuickFix but it is not legal to start and stop threads from within the initialisation and shutdown routines of a DLL (DllMain). You may not be doing this explicitly but QuickFix may be trying to clear up its threads when the DLL is unloaded and its static data is deleted automatically. The same effect does not happen when you put the Quickfix code into an EXE rather than a DLL - it is a frustrating limitation of Windows DLLs.
>
> Are you explicitly starting and stopping Quickfix, or do you assume that it will just clean itself up when your main EXE shuts down? If you are not doing so already, you might try creating, initializing and starting the Quickfix objects (Application etc) with an explicit call to the constructor (new CMyApplication etc), and doing the same thing to stop(), shutdown, delete everything explicitly before you unload the DLL or exit your program.
>
> I do not know that this would solve the problem but I have found that threads and DLLs do not mix very happily and doing this might remove the cause of your deadlocks.
>
> Dan
>
>> Date: Fri, 26 Dec 2008 17:25:26 -0500
>> From: ael...@gm...
>> To: qui...@li...
>> Subject: [Quickfix-developers] WaitForSingleObject deadlock
>>
>> QuickFIX Documentation: http://www.quickfixengine.org/quickfix/doc/html/index.html
>> QuickFIX Support: http://www.quickfixengine.org/services.html
>>
>> while doing some improvements to one of our QF systems, i was running
>> into some deadlocks. one group of which seem to be related to the
>> Windows API call to WaitForSingleObject as a thread join mechanism.
>> however for some reason in this situation, which is unfortunately
>> fairly complicated as I am attempting to extend the functionality of
>> another application by creating a DLL and implementing call backs for
>> the Application to call into and using an API to call into of the
>> applications, which requires that I use the Visual Studio 6 C/C++
>> compiler to create this DLL, it seems this call is not behaving itself
>> where instead of returning successfully when the thread function
>> returns, it's never returning when called with an INFINITE for the
>> timeout and returns with a timeout error when specified with a
>> timeout, even though the thread in question is returning (logged just
>> before return happens and same exact set of codes works where instead
>> of calling WaitForSingleObject instead at the end of the thread
>> function i set a flag to true signifying completion of thread and then
>> check that flag with a threaded sleep instead of calling
>> WaitForSingleObject).
>>
>> The above solution of course was applied to my application side logic
>> where threading was used, however Quickfix is attempting to make use
>> of the same Windows API call in Utility.cpp in the thread_join
>> function which of course with the above situation is causing any calls
>> to thread_join to hang indefinitely. Does anyone have any ideas why
>> this could be happening and a proper way to solve it?
>>
>> ------------------------------------------------------------------------------
>> _______________________________________________
>> Quickfix-developers mailing list
>> Qui...@li...
>> https://lists.sourceforge.net/lists/listinfo/quickfix-developers
>
> ________________________________
> Take your friends with you with Mobile Messenger. Click Here!
> ------------------------------------------------------------------------------
>
> _______________________________________________
> Quickfix-developers mailing list
> Qui...@li...
> https://lists.sourceforge.net/lists/listinfo/quickfix-developers
>

Re: [Quickfix-developers] WaitForSingleObject deadlock

[Peter P. <pet...@ho...>]

I don't know about the internals of QuickFix but it is not legal to start and stop threads from within the initialisation and shutdown routines of a DLL (DllMain). You may not be doing this explicitly but QuickFix may be trying to clear up its threads when the DLL is unloaded and its static data is deleted automatically. The same effect does not happen when you put the Quickfix code into an EXE rather than a DLL - it is a frustrating limitation of Windows DLLs.

Are you explicitly starting and stopping Quickfix, or do you assume that it will just clean itself up when your main EXE shuts down? If you are not doing so already, you might try creating, initializing and starting the Quickfix objects (Application etc) with an explicit call to the constructor (new CMyApplication etc), and doing the same thing to stop(), shutdown, delete everything explicitly before you unload the DLL or exit your program.

I do not know that this would solve the problem but I have found that threads and DLLs do not mix very happily and doing this might remove the cause of your deadlocks.

Dan

> Date: Fri, 26 Dec 2008 17:25:26 -0500
> From: ael...@gm...
> To: qui...@li...
> Subject: [Quickfix-developers] WaitForSingleObject deadlock
> 
> QuickFIX Documentation: http://www.quickfixengine.org/quickfix/doc/html/index.html
> QuickFIX Support: http://www.quickfixengine.org/services.html
> 
> while doing some improvements to one of our QF systems, i was running
> into some deadlocks.  one group of which seem to be related to the
> Windows API call to WaitForSingleObject as a thread join mechanism.
> however for some reason in this situation, which is unfortunately
> fairly complicated as I am attempting to extend the functionality of
> another application by creating a DLL and implementing call backs for
> the Application to call into and using an API to call into of the
> applications, which requires that I use the Visual Studio 6 C/C++
> compiler to create this DLL, it seems this call is not behaving itself
> where instead of returning successfully when the thread function
> returns, it's never returning when called with an INFINITE for the
> timeout and returns with a timeout error when specified with a
> timeout, even though the thread in question is returning (logged just
> before return happens and same exact set of codes works where instead
> of calling WaitForSingleObject instead at the end of the thread
> function i set a flag to true signifying completion of thread and then
> check that flag with a threaded sleep instead of calling
> WaitForSingleObject).
> 
> The above solution of course was applied to my application side logic
> where threading was used, however Quickfix is attempting to make use
> of the same Windows API call in Utility.cpp in the thread_join
> function which of course with the above situation is causing any calls
> to thread_join to hang indefinitely.  Does anyone have any ideas why
> this could be happening and a proper way to solve it?
> 
> ------------------------------------------------------------------------------
> _______________________________________________
> Quickfix-developers mailing list
> Qui...@li...
> https://lists.sourceforge.net/lists/listinfo/quickfix-developers

_________________________________________________________________
Are you a PC?  Upload your PC story and show the world 
http://clk.atdmt.com/UKM/go/122465942/direct/01/