On Tuesday 02 Mar 2010 20:54:18 Ralph Schindler wrote: > Hey all, > > From a code maintenance standpoint, the patch for this issue was > categorized as a "bug". This means the expected behavior was broken, > and is subject to being fixed in any release- Major, Minor, and even > Mini. Issue tracker improvements and feature requests can be included > in Minor releases, and Major proposal style features at Minor and Major > releases as well (Major including any BC breaks). > > I'd argue that your use case was outside of any of the intended use > cases for the class, strictly speaking: you were actually RELYING on > buggy behavior. > > Row objects should only contain information that the row is actually > modeling. That means that each row object should have the exact number > and type of fields that the database tables has that it originated from. > > The safest way of getting any intersection row objects (and data) would > be to use getDependentRowset() on that intersection table. This will > produce a working Row object that you can get the information from, and > update the data (if needed). > > I will discuss this with the team. > > -ralph
Hi Ralph, Thanks for your official reply. I'd like to hear the outcome. For me I can rewrite my code because it occurs only once. Others have bigger problems having multiple instances relying on this "bug". Regards, Jurian -- Jurian Sluiman CTO Soflomo V.O.F. http://soflomo.com
