API changes

[Patrick Linskey <pl...@gmail.com>]

Hi,

I think that I'm mostly done with the API changes -- see OPENJPA-317.

One outstanding issue is a naming problem. Internally, we use a
'FooMode' naming structure for lots of our symbolic constants, but the
JPA spec uses a 'FooType' naming structure for its enums. Which should
we obey? The most recent patch mostly goes the 'FooType' route.

-Patrick

-- 
Patrick Linskey
202 669 5907

Re: API changes

[Patrick Linskey <pl...@gmail.com>]

Hmm. Actually, I'll change things to be consistent, but use Type for
new things when it sounds nicer.

-Patrick

On 8/20/07, Patrick Linskey <pl...@gmail.com> wrote:
> OK, in that case, I'll switch things back to being Mode as appropriate.
>
> -Patrick
>
> On 8/20/07, Kevin Sutter <kw...@gmail.com> wrote:
> > I agree.  Let's be consistent.  If you go with FooType, then change
> > persistence.xml as well.  Otherwise, stick with FooMode.
> >
> > On 8/20/07, Patrick Linskey <pl...@gmail.com> wrote:
> > >
> > > Do you think that that means we should change the persistence.xml
> > > stuff, or go with the FooMode settings?
> > >
> > > It looks like the only ones that matter are EagerFetchMode,
> > > SubclassFetchMode, and ConnectionRetainMode. We also have
> > > ConnectionFactoryMode and TransactionMode, but these don't have
> > > corresponding enums, although there is still a consistency question.
> > >
> > > Also, for configuration settings, I'd be pretty happy just putting off
> > > the work until after 1.0. I don't think that consistency is as
> > > important there.
> > >
> > > -Patrick
> > >
> > > On 8/20/07, Marc Prud'hommeaux <mp...@apache.org> wrote:
> > > >
> > > > A agree with Kevin that I lean towards FooType, however I feel it is
> > > > more important to maintain consistency with the persistence.xml names
> > > > in cases where there is a choice to be made between FooType and FooMode.
> > > >
> > > >
> > > >
> > > > On Aug 20, 2007, at 7:49 PM, Patrick Linskey wrote:
> > > >
> > > > > So one issue with this is that some of these settings are configurable
> > > > > in persistence.xml, and we use 'FooMode' there. For example,
> > > > > ConnectionRetainMode.
> > > > >
> > > > > This is easy enough to fix, and can be done in the future by
> > > > > deprecating the current setting, so it's probably not a big
> > > > > consideration.
> > > > >
> > > > > -Patrick
> > > > >
> > > > > On 8/20/07, Kevin Sutter <kw...@gmail.com> wrote:
> > > > >> Patrick,
> > > > >> If I was forced to pick one, I would go with FooType, but I am
> > > > >> flexible
> > > > >> either way.
> > > > >>
> > > > >> Kevin
> > > > >>
> > > > >> On 8/20/07, Patrick Linskey <pl...@gmail.com> wrote:
> > > > >>>
> > > > >>> Hi,
> > > > >>>
> > > > >>> I think that I'm mostly done with the API changes -- see
> > > > >>> OPENJPA-317.
> > > > >>>
> > > > >>> One outstanding issue is a naming problem. Internally, we use a
> > > > >>> 'FooMode' naming structure for lots of our symbolic constants,
> > > > >>> but the
> > > > >>> JPA spec uses a 'FooType' naming structure for its enums. Which
> > > > >>> should
> > > > >>> we obey? The most recent patch mostly goes the 'FooType' route.
> > > > >>>
> > > > >>> -Patrick
> > > > >>>
> > > > >>> --
> > > > >>> Patrick Linskey
> > > > >>> 202 669 5907
> > > > >>>
> > > > >>
> > > > >
> > > > >
> > > > > --
> > > > > Patrick Linskey
> > > > > 202 669 5907
> > > >
> > > >
> > >
> > >
> > > --
> > > Patrick Linskey
> > > 202 669 5907
> > >
> >
>
>
> --
> Patrick Linskey
> 202 669 5907
>


-- 
Patrick Linskey
202 669 5907

Re: API changes

[Patrick Linskey <pl...@gmail.com>]

OK, in that case, I'll switch things back to being Mode as appropriate.

-Patrick

On 8/20/07, Kevin Sutter <kw...@gmail.com> wrote:
> I agree.  Let's be consistent.  If you go with FooType, then change
> persistence.xml as well.  Otherwise, stick with FooMode.
>
> On 8/20/07, Patrick Linskey <pl...@gmail.com> wrote:
> >
> > Do you think that that means we should change the persistence.xml
> > stuff, or go with the FooMode settings?
> >
> > It looks like the only ones that matter are EagerFetchMode,
> > SubclassFetchMode, and ConnectionRetainMode. We also have
> > ConnectionFactoryMode and TransactionMode, but these don't have
> > corresponding enums, although there is still a consistency question.
> >
> > Also, for configuration settings, I'd be pretty happy just putting off
> > the work until after 1.0. I don't think that consistency is as
> > important there.
> >
> > -Patrick
> >
> > On 8/20/07, Marc Prud'hommeaux <mp...@apache.org> wrote:
> > >
> > > A agree with Kevin that I lean towards FooType, however I feel it is
> > > more important to maintain consistency with the persistence.xml names
> > > in cases where there is a choice to be made between FooType and FooMode.
> > >
> > >
> > >
> > > On Aug 20, 2007, at 7:49 PM, Patrick Linskey wrote:
> > >
> > > > So one issue with this is that some of these settings are configurable
> > > > in persistence.xml, and we use 'FooMode' there. For example,
> > > > ConnectionRetainMode.
> > > >
> > > > This is easy enough to fix, and can be done in the future by
> > > > deprecating the current setting, so it's probably not a big
> > > > consideration.
> > > >
> > > > -Patrick
> > > >
> > > > On 8/20/07, Kevin Sutter <kw...@gmail.com> wrote:
> > > >> Patrick,
> > > >> If I was forced to pick one, I would go with FooType, but I am
> > > >> flexible
> > > >> either way.
> > > >>
> > > >> Kevin
> > > >>
> > > >> On 8/20/07, Patrick Linskey <pl...@gmail.com> wrote:
> > > >>>
> > > >>> Hi,
> > > >>>
> > > >>> I think that I'm mostly done with the API changes -- see
> > > >>> OPENJPA-317.
> > > >>>
> > > >>> One outstanding issue is a naming problem. Internally, we use a
> > > >>> 'FooMode' naming structure for lots of our symbolic constants,
> > > >>> but the
> > > >>> JPA spec uses a 'FooType' naming structure for its enums. Which
> > > >>> should
> > > >>> we obey? The most recent patch mostly goes the 'FooType' route.
> > > >>>
> > > >>> -Patrick
> > > >>>
> > > >>> --
> > > >>> Patrick Linskey
> > > >>> 202 669 5907
> > > >>>
> > > >>
> > > >
> > > >
> > > > --
> > > > Patrick Linskey
> > > > 202 669 5907
> > >
> > >
> >
> >
> > --
> > Patrick Linskey
> > 202 669 5907
> >
>


-- 
Patrick Linskey
202 669 5907

Re: API changes

[Kevin Sutter <kw...@gmail.com>]

I agree.  Let's be consistent.  If you go with FooType, then change
persistence.xml as well.  Otherwise, stick with FooMode.

On 8/20/07, Patrick Linskey <pl...@gmail.com> wrote:
>
> Do you think that that means we should change the persistence.xml
> stuff, or go with the FooMode settings?
>
> It looks like the only ones that matter are EagerFetchMode,
> SubclassFetchMode, and ConnectionRetainMode. We also have
> ConnectionFactoryMode and TransactionMode, but these don't have
> corresponding enums, although there is still a consistency question.
>
> Also, for configuration settings, I'd be pretty happy just putting off
> the work until after 1.0. I don't think that consistency is as
> important there.
>
> -Patrick
>
> On 8/20/07, Marc Prud'hommeaux <mp...@apache.org> wrote:
> >
> > A agree with Kevin that I lean towards FooType, however I feel it is
> > more important to maintain consistency with the persistence.xml names
> > in cases where there is a choice to be made between FooType and FooMode.
> >
> >
> >
> > On Aug 20, 2007, at 7:49 PM, Patrick Linskey wrote:
> >
> > > So one issue with this is that some of these settings are configurable
> > > in persistence.xml, and we use 'FooMode' there. For example,
> > > ConnectionRetainMode.
> > >
> > > This is easy enough to fix, and can be done in the future by
> > > deprecating the current setting, so it's probably not a big
> > > consideration.
> > >
> > > -Patrick
> > >
> > > On 8/20/07, Kevin Sutter <kw...@gmail.com> wrote:
> > >> Patrick,
> > >> If I was forced to pick one, I would go with FooType, but I am
> > >> flexible
> > >> either way.
> > >>
> > >> Kevin
> > >>
> > >> On 8/20/07, Patrick Linskey <pl...@gmail.com> wrote:
> > >>>
> > >>> Hi,
> > >>>
> > >>> I think that I'm mostly done with the API changes -- see
> > >>> OPENJPA-317.
> > >>>
> > >>> One outstanding issue is a naming problem. Internally, we use a
> > >>> 'FooMode' naming structure for lots of our symbolic constants,
> > >>> but the
> > >>> JPA spec uses a 'FooType' naming structure for its enums. Which
> > >>> should
> > >>> we obey? The most recent patch mostly goes the 'FooType' route.
> > >>>
> > >>> -Patrick
> > >>>
> > >>> --
> > >>> Patrick Linskey
> > >>> 202 669 5907
> > >>>
> > >>
> > >
> > >
> > > --
> > > Patrick Linskey
> > > 202 669 5907
> >
> >
>
>
> --
> Patrick Linskey
> 202 669 5907
>

Re: API changes

[Patrick Linskey <pl...@gmail.com>]

Do you think that that means we should change the persistence.xml
stuff, or go with the FooMode settings?

It looks like the only ones that matter are EagerFetchMode,
SubclassFetchMode, and ConnectionRetainMode. We also have
ConnectionFactoryMode and TransactionMode, but these don't have
corresponding enums, although there is still a consistency question.

Also, for configuration settings, I'd be pretty happy just putting off
the work until after 1.0. I don't think that consistency is as
important there.

-Patrick

On 8/20/07, Marc Prud'hommeaux <mp...@apache.org> wrote:
>
> A agree with Kevin that I lean towards FooType, however I feel it is
> more important to maintain consistency with the persistence.xml names
> in cases where there is a choice to be made between FooType and FooMode.
>
>
>
> On Aug 20, 2007, at 7:49 PM, Patrick Linskey wrote:
>
> > So one issue with this is that some of these settings are configurable
> > in persistence.xml, and we use 'FooMode' there. For example,
> > ConnectionRetainMode.
> >
> > This is easy enough to fix, and can be done in the future by
> > deprecating the current setting, so it's probably not a big
> > consideration.
> >
> > -Patrick
> >
> > On 8/20/07, Kevin Sutter <kw...@gmail.com> wrote:
> >> Patrick,
> >> If I was forced to pick one, I would go with FooType, but I am
> >> flexible
> >> either way.
> >>
> >> Kevin
> >>
> >> On 8/20/07, Patrick Linskey <pl...@gmail.com> wrote:
> >>>
> >>> Hi,
> >>>
> >>> I think that I'm mostly done with the API changes -- see
> >>> OPENJPA-317.
> >>>
> >>> One outstanding issue is a naming problem. Internally, we use a
> >>> 'FooMode' naming structure for lots of our symbolic constants,
> >>> but the
> >>> JPA spec uses a 'FooType' naming structure for its enums. Which
> >>> should
> >>> we obey? The most recent patch mostly goes the 'FooType' route.
> >>>
> >>> -Patrick
> >>>
> >>> --
> >>> Patrick Linskey
> >>> 202 669 5907
> >>>
> >>
> >
> >
> > --
> > Patrick Linskey
> > 202 669 5907
>
>


-- 
Patrick Linskey
202 669 5907

Re: API changes

[Marc Prud'hommeaux <mp...@apache.org>]

A agree with Kevin that I lean towards FooType, however I feel it is  
more important to maintain consistency with the persistence.xml names  
in cases where there is a choice to be made between FooType and FooMode.



On Aug 20, 2007, at 7:49 PM, Patrick Linskey wrote:

> So one issue with this is that some of these settings are configurable
> in persistence.xml, and we use 'FooMode' there. For example,
> ConnectionRetainMode.
>
> This is easy enough to fix, and can be done in the future by
> deprecating the current setting, so it's probably not a big
> consideration.
>
> -Patrick
>
> On 8/20/07, Kevin Sutter <kw...@gmail.com> wrote:
>> Patrick,
>> If I was forced to pick one, I would go with FooType, but I am  
>> flexible
>> either way.
>>
>> Kevin
>>
>> On 8/20/07, Patrick Linskey <pl...@gmail.com> wrote:
>>>
>>> Hi,
>>>
>>> I think that I'm mostly done with the API changes -- see  
>>> OPENJPA-317.
>>>
>>> One outstanding issue is a naming problem. Internally, we use a
>>> 'FooMode' naming structure for lots of our symbolic constants,  
>>> but the
>>> JPA spec uses a 'FooType' naming structure for its enums. Which  
>>> should
>>> we obey? The most recent patch mostly goes the 'FooType' route.
>>>
>>> -Patrick
>>>
>>> --
>>> Patrick Linskey
>>> 202 669 5907
>>>
>>
>
>
> -- 
> Patrick Linskey
> 202 669 5907

Re: API changes

[Patrick Linskey <pl...@gmail.com>]

So one issue with this is that some of these settings are configurable
in persistence.xml, and we use 'FooMode' there. For example,
ConnectionRetainMode.

This is easy enough to fix, and can be done in the future by
deprecating the current setting, so it's probably not a big
consideration.

-Patrick

On 8/20/07, Kevin Sutter <kw...@gmail.com> wrote:
> Patrick,
> If I was forced to pick one, I would go with FooType, but I am flexible
> either way.
>
> Kevin
>
> On 8/20/07, Patrick Linskey <pl...@gmail.com> wrote:
> >
> > Hi,
> >
> > I think that I'm mostly done with the API changes -- see OPENJPA-317.
> >
> > One outstanding issue is a naming problem. Internally, we use a
> > 'FooMode' naming structure for lots of our symbolic constants, but the
> > JPA spec uses a 'FooType' naming structure for its enums. Which should
> > we obey? The most recent patch mostly goes the 'FooType' route.
> >
> > -Patrick
> >
> > --
> > Patrick Linskey
> > 202 669 5907
> >
>


-- 
Patrick Linskey
202 669 5907

Re: API changes

[Kevin Sutter <kw...@gmail.com>]

Patrick,
If I was forced to pick one, I would go with FooType, but I am flexible
either way.

Kevin

On 8/20/07, Patrick Linskey <pl...@gmail.com> wrote:
>
> Hi,
>
> I think that I'm mostly done with the API changes -- see OPENJPA-317.
>
> One outstanding issue is a naming problem. Internally, we use a
> 'FooMode' naming structure for lots of our symbolic constants, but the
> JPA spec uses a 'FooType' naming structure for its enums. Which should
> we obey? The most recent patch mostly goes the 'FooType' route.
>
> -Patrick
>
> --
> Patrick Linskey
> 202 669 5907
>