Hi,
Ok, I've given CVS another spin and this bug still exists. It's actually not related to the boolean get/set like I thought it was, perhaps its morphed to a different bug after this time :)
Can someone tell me whether this is legitimately a bug? My guess is because I am setting with java.util.Collection, not ArrayList, but it works under 0.9.3, so I'm not sure if this was an original "bug", or a regression in CVS.
Will file in bugzilla with low priority...
Cheers,
Brett
...snip...
<field name="products" type="au.com.f2.shopping.product.Product" collection="arraylist" lazy="true">
<sql many-key="retailerId"/>
<bind-xml name="product" />
</field>
...snip...
/** Setter for property products.
* @param products New value of property products.
*/
public void setProducts(Collection products) {
this.products = products;
}
...snip...
org.exolab.castor.mapping.MappingException: The method set/addProducts in class au.com.f2.shopping.retailer.Retailer accepting/returning object of type class
au.com.f2.shopping.product.Product was not found
at org.exolab.castor.persist.FieldMolder.<init>(FieldMolder.java:584)
at org.exolab.castor.persist.ClassMolder.<init>(ClassMolder.java:367)
at org.exolab.castor.persist.ClassMolder.resolve(ClassMolder.java:519)
at org.exolab.castor.persist.LockEngine.<init>(LockEngine.java:168)
at org.exolab.castor.persist.PersistenceEngineFactory.createEngine(PersistenceEngineFactory.java:83)
at org.exolab.castor.jdo.engine.DatabaseRegistry.<init>(DatabaseRegistry.java:204)
at org.exolab.castor.jdo.engine.DatabaseRegistry.<init>(DatabaseRegistry.java:184)
at org.exolab.castor.jdo.engine.DatabaseRegistry.loadDatabase(DatabaseRegistry.java:317)
at org.exolab.castor.jdo.JDO.getDatabase(JDO.java:583)
-----Original Message-----
From: Brett Porter
Sent: Friday, 26 April 2002 4:00 PM
To: [EMAIL PROTECTED]
Subject: Re: [castor-dev] mapping boolean primitive
I use castor-0.9.3 and to use isXXX I do the get-method="isXXX", set-method="setXXX" thing, which works just fine.
However, I reported this being broken in CVS late last year, and it was still the case in 0.9.3.9.
I haven't checked CVS lately... after getting 0.9.3 working nicely for the most part I'm staying with the devil I know :) Once 0.9.4 makes it out I'll give it another spin, but mentioning "CVS version" to the tech. manager when discussing a production system tends cause deep psychological pain :)
Hope this clears it up - I might see if I can take the CVS version for a spin later today and see if that problem still exists.
Cheers,
Brett
