Re: [Gnumed-devel] gmObject problem

[Horst Herb]

On Sat, 26 Oct 2002 14:21, Horst Herb wrote:
> The big problem is:
> If such an object would be part of an outer transaction (invisible to this

Possible solution:
Instead of always passing a connection broker to the constructor of pgobject, 
we could pass a tuple (open read-only connection, open read-write connection) 
to the constructor.
Then, by checking the type of the parameter, the object could initialize 
itself accordingly.
If it is part of a larger outer transaction, the open write connection is used 
and thus final comittment would depend on the outer transaction.

Sounds simple enough to implement, can anybody see any major drawbacks (other 
than needing to know the transactional level at the time we create an 
object)?

Horst