James Procter wrote:
Ralf Joachim wrote:
your enhancement to hash code calculation sounds reasonable but even
if it is a very small addition we like to track that with jira. Can
you therefore please create a new issue in jira and attach all
relevant files.
many thanks. I thought that might be the response! I'll get round to it
as soon as I can.
I still haven't got around to formalising this as a JIRA bug fix and
test case, but have been testing the fix in place and have run into a
significant issue with cyclic references.
Are you aware of any existing mechanism within the castor source
generator for ensuring a method that recurses on the unmarshalled DOM
breaks cycles due to cyclic IDREF attributes :
<me id="1">
<bar id="3" meref="1"/>
</me>
This is a contrived example (obviously) - but you can see that with the
hashCode computation already implemented in castor, cycles will occur.
I have a way to fix this - but I'd rather use any existing mechanisms
for this kind of case if they already exist. Do they ?
many thanks,
Jim Procter.
--
-------------------------------------------------------------------
J. B. Procter (VAMSAS Project) Barton Bioinformatics Research Group
Phone/Fax:+44(0)1382 388734/345764 http://www.compbio.dundee.ac.uk
---------------------------------------------------------------------
To unsubscribe from this list please visit:
http://xircles.codehaus.org/manage_email
