RE: A couple of OSUser questions

I didn't make completely clear that I am a recent Hibernate convert.  That's
why I'm only now picking up the new OSUser code.  So you'll have to excuse my
current ignorance with composite keys under Hibernate.  Konstantin checked in
some example code into XDoclet samples for working with composites, but I
haven't had time to play with it.  If I remember correctly, it's using an
object as the key.  Clearly, it can't be a primitive type.
 
From a legacy perspective, people converting their code (or their brains)
from CMP are going to be using objects for their PKs.  
 
From my particular perspective, I have an entire schema full of object PKs.
Setting the PK to zero for an unset PK doesn't get caught with an NPE when I
get tired.  When refactoring code (either manually or automagically), it's a
lot harder to do so when the type is primitive versus object since object
based code usually uses Object.equals() for comparison.  
 
But having said all that, I started out the conversation simply looking for a
rationale for using primitives instead of objects.  The moment one assumes...
(cue sound of screetching tires and crunching metal...)  Maybe there were
some benefits to having primitives.  Just trying to figure out what's up...
 
best,
 
-b

-----Original Message-----
From: Matthew E. Porter [mailto:[email protected]]
Sent: Saturday, December 06, 2003 3:32 PM
To: [email protected]
Subject: Re: [Opensymphony-developers] A couple of OSUser questions


Eric: 
You may be right. I was thinking we were talking about a composite id. So,
back to Brian- what were you talking about? 

Also, if referring to long vs. Long, we are talking about 2 lines of code
that check whether it is equal to 0 under saveUser and saveGroup. 


Cheers, 
matthew 

On Dec 6, 2003, at 2:21 PM, Eric Pugh wrote: 


I didn't realize we where talking about using <composite-id>, as I agree, we
shouldn't use it.  However, the magic of Hibernate is that if my user
requires it, then in my implementation I can use a <composite-id> as long as
my mapping loads instead of the default one. 
I think that all he meant was to use Long versus long as the id, which would
still have the same mapping... 
Eric 
-----Original Message----- 
From:Matthew E. Porter [mailto:[email protected]] 
Sent:Saturday, December 06, 2003 7:56 PM 
To:[email protected] 
Subject:Re: [Opensymphony-developers] A couple of OSUser questions 


As quoting from Hibernate Documentation: 



"There is an alternative

<composite-id>declaration to allow access to legacy data with composite keys.
We strongly discourage its use for anything else."





Can we please hold off applying the foreign PK patch until more people have
chimed in? 



Cheers, 

matthew 



On Dec 6, 2003, at 1:44 PM, Eric Pugh wrote: 



Brian, 



I actually lean towards the PK being an object as well. I prefer to be able 

to know that getting a PK which is null means it hasn't been saved, versus 

knowing the 0 means unsaved. Especially as 0 is a perfectly good ID value! 




Which bug report was yours? I'll look at your patch and apply it, do you 

just want to one patch for all your changes? 



Eric 



-----Original Message----- 

From: [email protected] 

[mailto:[email protected]]On 

Behalf Of 

Brian Topping 

Sent: Saturday, December 06, 2003 7:16 PM 

To: [email protected] 

Subject: [Opensymphony-developers] A couple of OSUser questions 




Hi all, 



Back from some extended time away from the keys. Doing some 

travel and 

carpentry and things of that nature. It's fun to be back. 

Hope all your 

holidays are going well. 



Just getting recent versions of the code down and was psyched to see 

Hibernate support in OSUser. I've been working a bit with it 

in the last 24 

hours and filed one bug on it last night. I have the fixes 

for it, but it 

will take longer for someone to use them than it will for 

someone with IDEA 

and a RW connection to CVS to do the changes over again. It 

took me about 30 

minutes on this side. 



The other one, I'm not so sure it's a bug or not, it smells a 

lot like it 

belongs in the space of "conventions" rather than bug. So I 

thought I would 

ask for some rationale. 



In the CMP world, primary keys are objects instead of 

primitives. I've been 

following that through to Hibernate since it looks like 

composite types 

require that as well. But BaseHibernateEntity defines the PK 

as a primitive 

long. 



Any comments on why this would be beneficial? Otherwise, my 

vote is to 

convert to object PKs and am happy to do the work on it. 



thanks, 



-b 




------------------------------------------------------- 

This SF.net email is sponsored by: SF.net Giveback Program. 

Does SourceForge.net help you be more productive? Does it 

help you create better code? SHARE THE LOVE, and help us help 

YOU! Click Here: http://sourceforge.net/donate/ 

_______________________________________________ 

Opensymphony-developers mailing list 

[email protected] 

https://lists.sourceforge.net/lists/listinfo/opensymphony-developers 

<winmail.dat>