[ovirt-devel] Implementing equals & hashCode methods

[Yevgeny Zaspitsky]

See my comments in your mail body.

----- Original Message -----
> From: "Yair Zaslavsky" <yzaslavs at redhat.com>
> To: "Yevgeny Zaspitsky" <yzaspits at redhat.com>
> Cc: devel at ovirt.org
> Sent: Sunday, May 25, 2014 4:09:11 PM
> Subject: Re: [ovirt-devel] Implementing equals & hashCode methods
> 
> 
> 
> ----- Original Message -----
> > From: "Yevgeny Zaspitsky" <yzaspits at redhat.com>
> > To: devel at ovirt.org
> > Sent: Sunday, May 25, 2014 3:33:31 PM
> > Subject: [ovirt-devel] Implementing equals & hashCode methods
> > 
> > Hi All,
> > 
> > Recently I reviewed a patch that adds a new business entity to the engine.
> > The entity class has the following members:
> > 
> >     * id
> >     * data center id
> >     * name
> >     * type
> >     * some other properties that do not belong to the entity key
> > 
> > The equals & hashCode methods were implemented in a way that include all
> > members.
> > I asked the patch author to change that, so it'll include only business key
> > (data center id, name and type), which define the entity uniqueness.
> > Also I found that many other business entities are implemented in a similar
> > way (include all class members in equals & hashCode).
> > 
> > I'm a new to oVirt, so I'd like to ask your opinion on the issue.
> > 
> > 
> >     1. Do you agree with my approach on equals & hashCode.
> >     2. If you agree with my approach in general, should we implement it in
> >     the new introduced code or should we adhere to the old convention even
> >     we do not agree with it?
> >     3. Should we re-factor the old code (it might be dangerous as we do
> >     have
> >     enough unit test coverage)?
> > 
> > Thanks in advance,
> > Yevgeny
> 
> I assume your idea came (at least from ) the way identify is defined at
> Hibernate.
> The advantages are obvious -
> a. shorter code
> b. let's time to compute
> 
> However, I fear this is may lead to some bugs if identities are not defined
> well.

We should strive to define our classes well. Actually that point refers to the design stage rather than to the implementation.

On other hand, if we put the objects (that include all members in their equals) in a HashSet in order to keep unique instances, we might have a bug there.

> 
> Regarding 3 - might be dangerous as we don't have enough unit test coverage,
> probably.
> Actually, IMHO the situation with DAO tests which uses equals is quite good.

Those classes are used all over engine code and even in the UI code, so changing the implementation in the existing code is very risky. Having DAO covered isn't enough IMHO.

> For hashCode - I assume you have a point

As you probably know hashCode implementation could not be (conceptually) different from the equals one.

> 
> 
> 
> > 
> > _______________________________________________
> > Devel mailing list
> > Devel at ovirt.org
> > http://lists.ovirt.org/mailman/listinfo/devel
>