> Hmm.. In your case I actually don't see why the problem occurs :-/.
> "getIsProcessingOnly()" and "setIsTranscriptionClient" seem like
> benign property accessors without side effects. Is this so?
haha - good point. yes, it's a derivation that crawls the object graph
... so, quite likely SELECT was being generated therein.
> Anyway I guess logically it should be a no-no to change object state
> when commit is in progress (not sure if this is what actually happens
> in your case?).
yes, i agree--it was very backhanded.
> However I still thought that users should be allowed to perform select
> during the validation, as an object may need extra data to validate
> itself. So I opened the bug report and already changed the code to run
> the validation outside the iterator.
exactly -- as much as changing object state is a no-no in
validateForSave() ... deeper validation still needs to be allowed.
thanks andrus!!
This archive was generated by hypermail 2.0.0 : Sun Jun 13 2004 - 23:20:30 EDT