You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@felix.apache.org by Guillaume Nodet <gn...@apache.org> on 2016/05/03 16:54:24 UTC
[SCR] Some refactoring on SCR code
I've raised FELIX-5243 about refactoring some SCR code but haven't had any
feedback.
The commits are visible at:
https://github.com/gnodet/felix/commits/FELIX-5243
Unless someone complains, and given it's only refactoring and should not
alter the code itself (and all tests pass), I'll push those changes later
this week.
Cheers,
Guillaume
Re: [SCR] Some refactoring on SCR code
Posted by Guillaume Nodet <gn...@apache.org>.
Ok, i'll go ahead and commit ...
2016-05-19 21:30 GMT+02:00 David Jencks <da...@yahoo.com.invalid>:
> Now I did :-)
>
> I’m basically OK with this. I have a couple minor questions that I think
> we can resolve later unless you want to respond now.
>
> 1. I think Coercions should stay in the same package as Annotations
> since this is basically an implementation of the forthcoming object
> converter spec. I didn’t look to see if there is an appropriate spec
> method for a simple object to integer conversion which I think is the only
> non Annotations use of Coercions. Maybe we should copy the current object
> converter interface(s) and use them.
>
There is one usage, which I think was the reason why I did not move
Coercions.
It's used by DependencyManager#getMinimumCardinality
> 2. I don’t quite see why you have more or less changed the methods taking
> a RawParemeter to something else. Could you explain? Having done this
> shouldn’t we make the old methods less public?
>
You mean the various invoke() methods in ReferenceMethod / ComponentMethod.
I think I realised some classes were really only instantiated from the new
inject package, so it made sense to move them along and abstract a bit
more. If you think it's less clear, we can change back that bit.
>
> 3. I think there might be one or two new inner classes that perhaps should
> be static classes? I see one in ComponentRegistry.
>
The one createComponentHolder() I suppose ? yes, it definitely should be
static.
> I guess now I’d like you to commit this soon since I’m changing how the
> circular reference tracking works a little bit and that’s going to overlap
> with this.
>
> IIUC you did this in large part so you can reuse some bits elsewhere. How
> are we going to keep appropriate isolation going forwards? (i.e. so I don’t
> break your cdi stuff …)
>
That's definitely a good point. I'll try to setup a test to ensure that
the needed packages can be used in isolation.
>
> thanks!
> david jencks
>
> > On May 19, 2016, at 4:21 AM, Guillaume Nodet <gn...@apache.org> wrote:
> >
> > Did you had a chance to look at it ?
> >
> > Guillaume
> >
> > Le mardi 3 mai 2016, Guillaume Nodet <gnodet@apache.org <mailto:
> gnodet@apache.org>> a écrit :
> >
> >>
> >>
> >> 2016-05-03 19:12 GMT+02:00 David Jencks <david_jencks@yahoo.com.invalid
> <ma...@yahoo.com.invalid>
> >> <javascript:_e(%7B%7D,'cvml','david_jencks@yahoo.com.invalid <mailto:
> david_jencks@yahoo.com.invalid>');>>:
> >>
> >>> Hi Guillaume,
> >>>
> >>> Somehow I missed the issue being filed. I’d appreciate it if you
> could
> >>> hold off committing this until I have affirmatively reviewed the
> proposal.
> >>> I’ll try to get to it soon, but can’t promise.
> >>>
> >>
> >> There's no rush. I'm currently working on a copy of it, but obviously,
> I'd
> >> rather use the official SCR code at some point.
> >>
> >> Thx
> >>
> >>
> >>
> >>>
> >>> thanks
> >>> david jencks
> >>>
> >>>> On May 3, 2016, at 7:54 AM, Guillaume Nodet <gnodet@apache.org
> >>> <javascript:_e(%7B%7D,'cvml','gnodet@apache.org <mailto:
> gnodet@apache.org>');>> wrote:
> >>>>
> >>>> I've raised FELIX-5243 about refactoring some SCR code but haven't had
> >>> any
> >>>> feedback.
> >>>> The commits are visible at:
> >>>> https://github.com/gnodet/felix/commits/FELIX-5243
> >>>> Unless someone complains, and given it's only refactoring and should
> not
> >>>> alter the code itself (and all tests pass), I'll push those changes
> >>> later
> >>>> this week.
> >>>>
> >>>> Cheers,
> >>>> Guillaume
> >>>
> >>>
> >>
> >>
> >> --
> >> ------------------------
> >> Guillaume Nodet
> >> ------------------------
> >> Red Hat, Open Source Integration
> >>
> >> Email: gnodet@redhat.com <ma...@redhat.com>
> >> <javascript:_e(%7B%7D,'cvml','gnodet@redhat.com <mailto:
> gnodet@redhat.com>');>
> >> Web: http://fusesource.com <http://fusesource.com/>
> >> Blog: http://gnodet.blogspot.com/ <http://gnodet.blogspot.com/>
> >>
> >>
> >
> > --
> > ------------------------
> > Guillaume Nodet
> > ------------------------
> > Red Hat, Open Source Integration
> >
> > Email: gnodet@redhat.com <ma...@redhat.com>
> > Web: http://fusesource.com <http://fusesource.com/>
> > Blog: http://gnodet.blogspot.com/ <http://gnodet.blogspot.com/>
>
--
------------------------
Guillaume Nodet
------------------------
Red Hat, Open Source Integration
Email: gnodet@redhat.com
Web: http://fusesource.com
Blog: http://gnodet.blogspot.com/
Re: [SCR] Some refactoring on SCR code
Posted by David Jencks <da...@yahoo.com.INVALID>.
Now I did :-)
I’m basically OK with this. I have a couple minor questions that I think we can resolve later unless you want to respond now.
1. I think Coercions should stay in the same package as Annotations since this is basically an implementation of the forthcoming object converter spec. I didn’t look to see if there is an appropriate spec method for a simple object to integer conversion which I think is the only non Annotations use of Coercions. Maybe we should copy the current object converter interface(s) and use them.
2. I don’t quite see why you have more or less changed the methods taking a RawParemeter to something else. Could you explain? Having done this shouldn’t we make the old methods less public?
3. I think there might be one or two new inner classes that perhaps should be static classes? I see one in ComponentRegistry.
I guess now I’d like you to commit this soon since I’m changing how the circular reference tracking works a little bit and that’s going to overlap with this.
IIUC you did this in large part so you can reuse some bits elsewhere. How are we going to keep appropriate isolation going forwards? (i.e. so I don’t break your cdi stuff …)
thanks!
david jencks
> On May 19, 2016, at 4:21 AM, Guillaume Nodet <gn...@apache.org> wrote:
>
> Did you had a chance to look at it ?
>
> Guillaume
>
> Le mardi 3 mai 2016, Guillaume Nodet <gnodet@apache.org <ma...@apache.org>> a écrit :
>
>>
>>
>> 2016-05-03 19:12 GMT+02:00 David Jencks <david_jencks@yahoo.com.invalid <ma...@yahoo.com.invalid>
>> <javascript:_e(%7B%7D,'cvml','david_jencks@yahoo.com.invalid <ma...@yahoo.com.invalid>');>>:
>>
>>> Hi Guillaume,
>>>
>>> Somehow I missed the issue being filed. I’d appreciate it if you could
>>> hold off committing this until I have affirmatively reviewed the proposal.
>>> I’ll try to get to it soon, but can’t promise.
>>>
>>
>> There's no rush. I'm currently working on a copy of it, but obviously, I'd
>> rather use the official SCR code at some point.
>>
>> Thx
>>
>>
>>
>>>
>>> thanks
>>> david jencks
>>>
>>>> On May 3, 2016, at 7:54 AM, Guillaume Nodet <gnodet@apache.org
>>> <javascript:_e(%7B%7D,'cvml','gnodet@apache.org <ma...@apache.org>');>> wrote:
>>>>
>>>> I've raised FELIX-5243 about refactoring some SCR code but haven't had
>>> any
>>>> feedback.
>>>> The commits are visible at:
>>>> https://github.com/gnodet/felix/commits/FELIX-5243
>>>> Unless someone complains, and given it's only refactoring and should not
>>>> alter the code itself (and all tests pass), I'll push those changes
>>> later
>>>> this week.
>>>>
>>>> Cheers,
>>>> Guillaume
>>>
>>>
>>
>>
>> --
>> ------------------------
>> Guillaume Nodet
>> ------------------------
>> Red Hat, Open Source Integration
>>
>> Email: gnodet@redhat.com <ma...@redhat.com>
>> <javascript:_e(%7B%7D,'cvml','gnodet@redhat.com <ma...@redhat.com>');>
>> Web: http://fusesource.com <http://fusesource.com/>
>> Blog: http://gnodet.blogspot.com/ <http://gnodet.blogspot.com/>
>>
>>
>
> --
> ------------------------
> Guillaume Nodet
> ------------------------
> Red Hat, Open Source Integration
>
> Email: gnodet@redhat.com <ma...@redhat.com>
> Web: http://fusesource.com <http://fusesource.com/>
> Blog: http://gnodet.blogspot.com/ <http://gnodet.blogspot.com/>
Re: [SCR] Some refactoring on SCR code
Posted by Guillaume Nodet <gn...@apache.org>.
Did you had a chance to look at it ?
Guillaume
Le mardi 3 mai 2016, Guillaume Nodet <gn...@apache.org> a écrit :
>
>
> 2016-05-03 19:12 GMT+02:00 David Jencks <david_jencks@yahoo.com.invalid
> <javascript:_e(%7B%7D,'cvml','david_jencks@yahoo.com.invalid');>>:
>
>> Hi Guillaume,
>>
>> Somehow I missed the issue being filed. I’d appreciate it if you could
>> hold off committing this until I have affirmatively reviewed the proposal.
>> I’ll try to get to it soon, but can’t promise.
>>
>
> There's no rush. I'm currently working on a copy of it, but obviously, I'd
> rather use the official SCR code at some point.
>
> Thx
>
>
>
>>
>> thanks
>> david jencks
>>
>> > On May 3, 2016, at 7:54 AM, Guillaume Nodet <gnodet@apache.org
>> <javascript:_e(%7B%7D,'cvml','gnodet@apache.org');>> wrote:
>> >
>> > I've raised FELIX-5243 about refactoring some SCR code but haven't had
>> any
>> > feedback.
>> > The commits are visible at:
>> > https://github.com/gnodet/felix/commits/FELIX-5243
>> > Unless someone complains, and given it's only refactoring and should not
>> > alter the code itself (and all tests pass), I'll push those changes
>> later
>> > this week.
>> >
>> > Cheers,
>> > Guillaume
>>
>>
>
>
> --
> ------------------------
> Guillaume Nodet
> ------------------------
> Red Hat, Open Source Integration
>
> Email: gnodet@redhat.com
> <javascript:_e(%7B%7D,'cvml','gnodet@redhat.com');>
> Web: http://fusesource.com
> Blog: http://gnodet.blogspot.com/
>
>
--
------------------------
Guillaume Nodet
------------------------
Red Hat, Open Source Integration
Email: gnodet@redhat.com
Web: http://fusesource.com
Blog: http://gnodet.blogspot.com/
Re: [SCR] Some refactoring on SCR code
Posted by Guillaume Nodet <gn...@apache.org>.
2016-05-03 19:12 GMT+02:00 David Jencks <da...@yahoo.com.invalid>:
> Hi Guillaume,
>
> Somehow I missed the issue being filed. I’d appreciate it if you could
> hold off committing this until I have affirmatively reviewed the proposal.
> I’ll try to get to it soon, but can’t promise.
>
There's no rush. I'm currently working on a copy of it, but obviously, I'd
rather use the official SCR code at some point.
Thx
>
> thanks
> david jencks
>
> > On May 3, 2016, at 7:54 AM, Guillaume Nodet <gn...@apache.org> wrote:
> >
> > I've raised FELIX-5243 about refactoring some SCR code but haven't had
> any
> > feedback.
> > The commits are visible at:
> > https://github.com/gnodet/felix/commits/FELIX-5243
> > Unless someone complains, and given it's only refactoring and should not
> > alter the code itself (and all tests pass), I'll push those changes later
> > this week.
> >
> > Cheers,
> > Guillaume
>
>
--
------------------------
Guillaume Nodet
------------------------
Red Hat, Open Source Integration
Email: gnodet@redhat.com
Web: http://fusesource.com
Blog: http://gnodet.blogspot.com/
Re: [SCR] Some refactoring on SCR code
Posted by David Jencks <da...@yahoo.com.INVALID>.
Hi Guillaume,
Somehow I missed the issue being filed. I’d appreciate it if you could hold off committing this until I have affirmatively reviewed the proposal. I’ll try to get to it soon, but can’t promise.
thanks
david jencks
> On May 3, 2016, at 7:54 AM, Guillaume Nodet <gn...@apache.org> wrote:
>
> I've raised FELIX-5243 about refactoring some SCR code but haven't had any
> feedback.
> The commits are visible at:
> https://github.com/gnodet/felix/commits/FELIX-5243
> Unless someone complains, and given it's only refactoring and should not
> alter the code itself (and all tests pass), I'll push those changes later
> this week.
>
> Cheers,
> Guillaume