Re: [Springframework-developer] AbstractLobType#nullSafeGetInternal((ResultSet, int, LobHandler) is missing a parameter...

[Dan W. <dan...@gm...>]

Hi All,
I've posted an updated implementation of FileType on the forum
(http://forum.springframework.org/viewtopic.php?p=3D17265#17265).  I'd
be interested in any feedback people might have.

I'm leaning toward making the temporary file name more unique by
adding a random number to the end of the owner.hashCode() value then
letting the user manage the caching of files..

Cheers,
Dan. =20


On Mon, 28 Mar 2005 15:24:37 +1000, Dan Washusen <dan...@gm...> w=
rote:
> Awesome, thanks Juergen...
>=20
> On Fri, 25 Mar 2005 09:43:53 +0100, Juergen Hoeller
> <ju...@in...> wrote:
> > FYI, I've extended the signature of AbstractLobType's "nullSafeGetInter=
nal"
> > to include the owner object and also the full array of column names as
> > passed in by Hibernate. The concrete subclasses that we ship don't use
> > those, so this is only meant for custom subclasses.
> >
> > I've only applied this change to the Hibernate3 support. I don't think =
that
> > it is necessary to perform an analogous change in the Hibernate2 suppor=
t.
> > After all, noone has asked for it before; I assume that new development=
s
> > will use Hibernate3 anyway.
> >
> > I've committed the change yesterday evening; this will make it into 1.2=
 RC1!
> >
> > Juergen
> >
> > -----Original Message-----
> > From: spr...@li...
> > [mailto:spr...@li...]On Behalf
> > Of Colin Sampaleanu
> > Sent: Wednesday, March 23, 2005 11:58 PM
> > To: spr...@li...
> > Subject: Re: [Springframework-developer]
> > AbstractLobType#nullSafeGetInternal((ResultSet, int, LobHandler) is
> > missing a parameter...
> >
> > I doubt there are many people subclassing it, and it would be trivial
> > for anybody making the change to fix-up their code for the difference.
> > They're going to be modifying their code anyway to account for package
> > changes.
> >
> > Colin
> >
> > Washusen, Dan wrote:
> >
> > > It would make the upgrade from Hibernate 2 to Hibernate 3 a little
> > > harder. Although, I wouldn't think many people have custom
> > > implementations of AbstractLobType=85
> > >
> > > ---------------------------------------------------------------------=
---
> > >
> > > *From:* spr...@li...
> > > [mailto:spr...@li...] *On
> > > Behalf Of *Juergen Hoeller
> > > *Sent:* Thursday, 24 March 2005 12:59 AM
> > > *To:* spr...@li...
> > > *Subject:* Re: [Springframework-developer]
> > > AbstractLobType#nullSafeGetInternal((ResultSet, int, LobHandler) is
> > > missing a parameter...
> > >
> > > Well, backwards compatibility just affects the Hibernate 2.x version.
> > > The Hibernate3 version was migrated straight from that. We could add
> > > the owner object in the Hibernate3 version; after all, we haven't
> > > officially release the Hibernate3 support yet, so are not constrained
> > > by backwards compatibility there.
> > >
> > > Juergen
> > >
> > >     -----Original Message-----
> > >     *From:* spr...@li...
> > >     [mailto:spr...@li...]*On
> > >     Behalf Of *Washusen, Dan
> > >     *Sent:* Tuesday, March 22, 2005 11:39 PM
> > >     *To:* spr...@li...
> > >     *Subject:* RE: [Springframework-developer]
> > >     AbstractLobType#nullSafeGetInternal((ResultSet, int, LobHandler)
> > >     is missing a parameter...
> > >
> > >     Hi Juergen,
> > >
> > >     The version I was looking at (hibernate3) has a since of 1.2. I
> > >     mistakenly presumed it was new; backward-compatibility is a very
> > >     good reason not to change it...
> > >
> > >     I was trying to make an implementation of AbstractLobType that
> > >     didn't read everything into memory. The code is available here
> > >     (http://forum.springframework.org/viewtopic.php?t=3D4237). I was
> > >     going to use the owner to generate a hash to use as a temporary
> > >     file name=85
> > >
> > >     I guess I'll re-implement the logic in AbstractLobType=85
> > >
> > >     Cheers,
> > >
> > >     Dan
> > >
> > >     -----------------------------------------------------------------=
-----
> > --
> > >
> > >     *From:* spr...@li...
> > >     [mailto:spr...@li...] *O=
n
> > >     Behalf Of *Juergen Hoeller
> > >     *Sent:* Tuesday, 22 March 2005 5:32 PM
> > >     *To:* spr...@li...
> > >     *Subject:* Re: [Springframework-developer]
> > >     AbstractLobType#nullSafeGetInternal((ResultSet, int, LobHandler)
> > >     is missing a parameter...
> > >
> > >     Hi Dan,
> > >
> > >     I guess there simply hasn't been a need to access the owner objec=
t
> > >     in one of the subclasses yet. AbstractLobType mainly serves as
> > >     base class for the concrete LOB types that we ship in Spring,
> > >     which don't need owner access.
> > >
> > >     The problem is that the "nullSafeGetInternal" method would have t=
o
> > >     change in a non-backwards-compatible fashion to expose the owner,
> > >     so I'm reluctant to change this without a compelling reason.
> > >
> > >     Juergen
> > >
> > >         -----Original Message-----
> > >         *From:* spr...@li...
> > >         [mailto:spr...@li...=
]*On
> > >         Behalf Of *Washusen, Dan
> > >         *Sent:* Tuesday, March 22, 2005 4:05 AM
> > >         *To:* spr...@li...
> > >         *Subject:* [Springframework-developer]
> > >         AbstractLobType#nullSafeGetInternal((ResultSet, int,
> > >         LobHandler) is missing a parameter...
> > >
> > >         Hey guys,
> > >
> > >         I'm trying to use the new
> > >         org.springframework.orm.hibernate3.support.AbstractLobType an=
d
> > >         I'm wondering why the nullSafeGet implementation doesn't pass
> > >         the owner parameter to the abstract nullSafeGetInternal metho=
d.
> > >
> > >         Should I raise a Jira issue?
> > >
> > >         Cheers,
> > >
> > >         Dan
> > >