I was just curious as to why it was decided to make getMapBuilder() static 
in BasePeer. It doesn't seem like it is really since the map builder is 
different from each generated class.

I do recognize that it allows the call to made from a static context and
that most of the methods in the generated peer are static. It is handy to
be able to call it, but it is not really a static method in BasePeer, it
is really abstract.  Having it be static makes it very difficult for
someone who is writing a superclass for a peer to get a transaction
because a call to getMapBuilder() (to get the database name) calls the
method in BasePeer rather than the subclass because static methods are not
inherited.

I am working on putting several different trees in a database and torque 
has been working very well to make it so that the defined classes will 
just work except in this one area.

In my opinion declaring getMapBuilder() static, though syntactically
correct, is not correct from a design perspective and should be fixed in 
the next major api change.

Will Holcomb


--
To unsubscribe, e-mail:   <mailto:turbine-torque-user-unsubscribe@;jakarta.apache.org>
For additional commands, e-mail: <mailto:turbine-torque-user-help@;jakarta.apache.org>

Reply via email to