Hi,

I've cloned the issue QFJ-403 cause I'm in the same case. My counterpart does not set the OrigSendingTime for dup message.
I don't see any solution, except they correct their soft or hacking quickfix/j.

QFJ-403 was marked fixed, so presumably this shouldn't be happening. Are you using an old version of QF/J?

Yes, the problem is marked fixed and I use the last 1.5.1 version (that include the patch).
But as I say, my counterpart does not set the OrigSendingTime for dup message (FIX 4.2) and I'm looking for a workaround (correct answer should be "they have to respect the protocol" but it seems to be difficult - big company, lots of customers).

The only way I found to intercept the message to add missing tag 122 is to set an IoFilterChainBuilder on SocketInitiator.
But the problem is the message received in interceptor method messageReceived() is a String, so I have to parse it.

Is there a better way to intercept FIX message before session handles it?

Maybe I should create a new ticket: "How can I intercept quickfix message before session handles it", no?

Hi guys,

indeed, QFJ-403 is marked as fixed. But the proposed solution was not incorporated into QF/J. At least the code in validatePossDup(Message) still looks like this (on 1.5.1 and on trunk):

if (!header.isSetField(OrigSendingTime.FIELD)) {
    generateReject(msg, SessionRejectReason.REQUIRED_TAG_MISSING, OrigSendingTime.FIELD);
    return false;
}

Hi Christophe,
I will take care of this. It seems the fix was not applied or got reverted.
Cheers,
Chris.

Hi,
Setting "RejectInvalidMessage" flag to "Y" is too risky when I want ignore ONLY 122 field but I want to validate rest of fields. I propose reopen this bug.

Maybe should be added property something like:
IgnoredFieldsDueValitation=<field1>,<field2>...
and errors in field1, field2 will be ignored, but not in the other fields?

P.S. Sorry for my english

Cześć Krzysztof,
I think we should just check if tag 122/OrigSendingTime is present and if it isn't we shouldn't do the SendingTime vs. OrigSendingTime check. In my opinion QF/J should rather try to synchronize correctly instead of throwing validation errors. -> QFJ-703

I agree. So I am waiting for new version. Thanks