On Wed, 2002-11-27 at 15:49, John McNally wrote:
> I'm guessing this was a result of hitting Send to quickly, did you want
> this to go to torque-dev?
> 

Yes you're correct, I didn't notice that scarab sets a different
reply-to.

> I will note that the retrieveByPK method is usable in its current
> state.  When I wrote the method, the intention was that you should not
> call it with a random number, but one that is known to be a the pk of
> some row.  But allowing its use as a way to determine if a certain value
> is a pk is reasonable I guess.

The problem is the traditional stateless nature of HTTP. The row in the
database did exist, but might have been deleted before we attempted to
access it.  This is a small error that the application should be able to
handle gracefully. The other extreme is the case where there are
multiple results, which would indicate a major failure on the database
softwares part. It seems to me that TorqueException traditionally
indicates a major failure that my application is not going to be able to
solve on it's own.

> I'm not sure why you disagree with calling the new method
> getByPrimaryKey.

Part of the reason I don't like that method name is that it is following
the bean pattern, even though it isn't returning a "ByPrimaryKey".  I
think it will also confuse developers... "Why is there a get and a
retrieve?"

I've added the getByPrimaryKey methods to my copy of Torque, so I can
upload the patch if people really want to do it this way.  It involves a
lot of duplicated code, someone else may want to modify retrieveByPK to
call getByPrimaryKey.  I don't really understand aliases, so I don't
understand what all that code is trying to accomplish.

-james

-- 
James A. Hillyerd <[EMAIL PROTECTED]>
Java Software Engineer - http://www.whynotown.com/


--
To unsubscribe, e-mail:   <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>

Reply via email to