Re: [Quickfixj-users] Parsing semantics?

[Christoph J. <chr...@ma...>]

Hi Robert,

if you want to suppress as much validation as possible you should use the following settings: (see 
explanation at https://www.quickfixj.org/usermanual/2.1.0/usage/configuration.html#Validation)
ValidateIncomingMessage = N
ValidateFieldsOutOfOrder = N
ValidateFieldsHaveValues =N
ValidateUserDefinedFields = N
ValidateUnorderedGroupFields = N
AllowUnknownMsgFields = Y

Keep in mind that you still need to use the data dictionary to parse your known repeating groups 
correctly. So in any case set UseDataDictionary=Y.

The latest changes around that part were done with release 1.6.4, so about three years ago:
https://www.quickfixj.org/jira/browse/QFJ-791
https://www.quickfixj.org/jira/browse/QFJ-169

Before these changes the resulting message was truncated on the first failing tag.

But still after these changes, if you have a whole unknown repeating group then you might end up 
with the tags scattered over the message (ordered by tag) since QFJ has no way of knowing that this 
is a repeating group.
But for only one new tag it should work so that the new tag will be placed into the repeating group.

Hope that helps.
Cheers,
Chris.


On 15.09.20 18:50, Robert Nicholson wrote:
> QuickFIX/J Documentation: http://www.quickfixj.org/documentation/
> QuickFIX/J Support: http://www.quickfixj.org/support/
>
>
> I have RejectInvalidMessage=N in order to avoid a vendor adding new tags or new enum values breaking the processing of a message on my side if it’s a tag that I don’t care about.
>
> However, it’s evident that in earlier releases of quickfixj that this doesn’t insulate you from this kind of event because it will effectively stop parsing the message as soon as it finds something it doesn’t like. Even if it’s just going to warn about it.
>
> If somebody adds a new repeating group tag in the message that my FIX dictionary knows nothing about then I’ll still process the message but it’s extremely unlikely that it will parse the content after
> the new repeating group therefore resulting in a partially parsed message being processed.
>
> Q. is this the correct understanding and do all quickfixj releases work this way?
>
> If not all releases work this way when did this behavior change such that it makes a best effort at parsing the message?
>
> What settings allow me to still be able to fully parse the message ignoring any new content that doesn’t align with my dictionary but still end up with a message equivalent to that which didn’t have the new repeating group tags?
>
> _______________________________________________
> Quickfixj-users mailing list
> Qui...@li...
> https://lists.sourceforge.net/lists/listinfo/quickfixj-users

-- 
Christoph John
Software Engineering
T +49 241 557080-28
chr...@ma...

MACD GmbH
Oppenhoffallee 103
52066 Aachen, Germany
www.macd.com

Amtsgericht Aachen: HRB 8151
Ust.-Id: DE 813021663
Geschäftsführer: George Macdonald