Re: Constructors even more crippled in 0.43?

• To: Ozone DB Mailing List <ozone-users@ozone-db.org>
• Subject: Re: Constructors even more crippled in 0.43?
• From: Falko Braeutigam <falko@softwarebuero.de>
• Date: Tue, 15 Feb 2000 17:22:45 +0100
• In-Reply-To: <1129416.3159541579@turtle.coral.cs.cmu.edu>
• Organization: SMB
• References: <1129416.3159541579@turtle.coral.cs.cmu.edu>

On Tue, 15 Feb 2000, William Uther wrote:
> Hi,
>   It seems that constructors have been further crippled in 0.4dev3.
> Previously self() didn't work in constructors, now database() doesn't work.
> This means you can't make a new object in an object constructor.  Sigh.

It's not that easy to work around this limitation. The container of the
object calls setContainer() (which makes self() and database() work) just after
the object is created. In fact this is to late for the constructor. But on the
other hand I need an object on which I can call setContainer(). To solve this
the constructor itself has to find out its associated container. This would
bring back database dependent code into OzoneObject ctor which in turn makes it
again more difficult to implement OzoneCompatible on class trees that already
exists. What to do?


Falko
-- 
______________________________________________________________________
Falko Braeutigam                         mailto:falko@softwarebuero.de
softwarebuero m&b (SMB)                    http://www.softwarebuero.de

• References:
  • Constructors even more crippled in 0.43?
    • From: William Uther <will+ozone-db@cs.cmu.edu>

• Prev by Date: Re: DxHashMap can't clone when empty
• Next by Date: Object creation is slow. Disk Access?
• Prev by thread: Constructors even more crippled in 0.43?
• Next by thread: Object creation is slow. Disk Access?
• Index(es):
  • Date
  • Thread