Re: User subclassing

2006-07-23 Thread Kenneth Gonsalves
On 22-Jul-06, at 11:04 PM, Todd O'Bryan wrote: >> >> Again, having the solution with OneToOne relation scares newbies away >> because it's become a kind of common wisdom that OneToOne will be >> changed very drastically very soon and "all your code would burn". > > What is the current wisdom on

Re: User subclassing

2006-07-22 Thread James Bennett
On 7/22/06, Todd O'Bryan <[EMAIL PROTECTED]> wrote: > Underlying behavior? So, is it basically safe to use a OneToOneField > and the API, though not guaranteed to be consistent, is likely to > just require tweaking? In other words, while underlying behavior > might change, it'll look roughly equiv

Re: User subclassing

2006-07-22 Thread Todd O'Bryan
On Jul 22, 2006, at 7:18 PM, James Bennett wrote: > > On 7/22/06, Todd O'Bryan <[EMAIL PROTECTED]> wrote: >> What is the current wisdom on OneToOne? Will it be finalized soon? >> Ever? > > As I understand it from a conversation I had the other day, nothing > whatsoever will happen until subcla

Re: User subclassing

2006-07-22 Thread Bryan
The only framework that I've seen sufficiently tackle this whole user extension problem is asp.net 2.0 with their provider model. The provider model allows users to extend or "plug" in their own implementation of a feature. This is useful when the given schema doesn't meet your needs, such as a

Re: User subclassing

2006-07-22 Thread James Bennett
On 7/22/06, Todd O'Bryan <[EMAIL PROTECTED]> wrote: > What is the current wisdom on OneToOne? Will it be finalized soon? Ever? As I understand it from a conversation I had the other day, nothing whatsoever will happen until subclassing is straightened out. At that point we'll move to recommending

Re: User subclassing

2006-07-22 Thread Todd O'Bryan
On Jul 22, 2006, at 2:59 AM, Ivan Sagalaev wrote: > > Again, having the solution with OneToOne relation scares newbies away > because it's become a kind of common wisdom that OneToOne will be > changed very drastically very soon and "all your code would burn". What is the current wisdom on OneToO

User subclassing

2006-07-21 Thread Ivan Sagalaev
James Bennett wrote: > Which I've never understood. Subclassing User, even when model > inheritance works again, will be probably the most difficult, least > efficient and least rewarding way of extending User... But nobody knows this beforehand. The feature is called "subclassing" and looks lik