Hi!

On 01/16/2008 03:07 PM Frederik Holljen wrote:
>>> No, I don't think we can find a way here, either. We would still need to
>>> introduce a new definition class to be used alternatively to
>>> ezcPersistentObjectProperty (the current defintion class for
>>> properties). This would mean checks anywhere such a property is used in
>>> a query, which applies to as many places as it does for the ID property.

> A better way is to create a whole new definition class that supports
> both the current and the new id's and to have our code support only
> that. In the definition loader we detect if there are old style
> definitions which we simply convert to new style definitions. This way
> we don't make the code more complicated than necessary.

I think this would be step for version 2.0 of PersistentObject. Doing 
this now isn't a good idea, since the conversion will take quite some 
time on load. Beside that, it requires a completly new design for 
defintion files as well as a re-implementation of all currently 
supported features.

>>> Therefore I'd suggest to close #8963 as "Won't implement".

> Although I think this issue is hard and complicated I think we should
> support it eventually since not having this functionality can be a
> showstopper for anyone with legacy databases.

This is true for composite IDs, yes. We should re-consider this if we 
really get the requirement to implement it.

Regards,
Toby
-- 
Mit freundlichen Grüßen / Med vennlig hilsen / With kind regards

Tobias Schlitt (GPG: 0xC462BC14) eZ Components Developer

[EMAIL PROTECTED] | eZ Systems AS | ez.no
-- 
Components mailing list
[email protected]
http://lists.ez.no/mailman/listinfo/components

Reply via email to