Hi,

I'm +1 for fixing this issue in QFJ side. I feel that losing CDATA sections
in the infoset along with the binary support is a too much of a price to
pay.

On Sat, Aug 16, 2008 at 1:40 AM, Asanka Abeysinghe <[EMAIL PROTECTED]> wrote:

>  Hi,
> I did fix the issue QFJ-330 and provided the patch. The change is to make
> one package access method to public access inside
> core/src/main/java/quickfix/FieldMap.java. This method returns a list of
> groups (List<Group> getGroups(int field)). And the group contains the 
> ***delimiter
> field***. Purpose of accessing the DD is to get the delimiter key for the
> repeating group, with this approach we don't need to access the DD.
>

In that case I don't think there will be any performance degradations. Am I
correct?


>
> With this change current implementation of FIX to AXIOM will remain with
> CDATA and binary support it will add new routines to create and parse the
> repeating groups using recursions.
>

+1


>
> Do we have to wait till QFJ accept QFJ-330 patch? if not we have to deploy
> the updated quickfixj-core-<version>.jar in a public maven repo and provide
> a download url for the synapse FIX users to download the updated library.
>

I don't think QFJ folks would accept the patch until their next release. I
submitted a patch few months back and it wasn't accepted until their latest
release. So I think it's best to look for other options like hosting the
modified library in a public Maven repo.

Thanks

Best Regards,
Hiranya


> Asanka A.
>
>
>
>
>
>

Reply via email to