Re: [convert] API Discussion...

2011-11-05 Thread Simone Tripodi
Yes, there is space for a lot of improvements, in both therms of APIs and performances! :) http://people.apache.org/~simonetripodi/ http://simonetripodi.livejournal.com/ http://twitter.com/simonetripodi http://www.99soft.org/ On Sat, Nov 5, 2011 at 4:27 PM, James Carman wrote: > Well, with tha

Re: [convert] API Discussion...

2011-11-05 Thread James Carman
Well, with that idea, it got me interested in Meiyo. So, now I'm playing around in a DSL branch trying to improve the API. It seems too verbose to me at the moment. :) On Sat, Nov 5, 2011 at 11:21 AM, Simone Tripodi wrote: > Hi James! > I like the idea, the concept reminds me how TestNG's Data

Re: [convert] API Discussion...

2011-11-05 Thread Simone Tripodi
Hi James! I like the idea, the concept reminds me how TestNG's DataProvider and GoogleGuice Providers. Looking forward to see it in action! Simo http://people.apache.org/~simonetripodi/ http://simonetripodi.livejournal.com/ http://twitter.com/simonetripodi http://www.99soft.org/ On Sat, Nov 5,

Re: [convert] API Discussion...

2011-11-05 Thread James Carman
What if we introduce a @Converter annotation and any method that is annotated with this annotation is automatically registered as a converter? It's similar to what I've done in Metastopheles (http://metastopheles.sourceforge.net/). On Fri, Nov 4, 2011 at 8:15 AM, Adrian Crum wrote: > Agreed. Ple

Re: [convert] API Discussion...

2011-11-04 Thread Simone Tripodi
I am interested on see Jame's code, a brach would work perfectly for me! Matt Benson has nice stuff too at his hands, it would be great having all of them merged in one project. Maybe one day [beanutils] will be the main [convert] consumer... :) Simo http://people.apache.org/~simonetripodi/ http:/

Re: [convert] API Discussion...

2011-11-04 Thread Adrian Crum
Agreed. Please don't mis-interpret my replies - I'm not trying to "own" the sandbox, I just want everyone to have a chance to play in it. The recent interest in Convert is great - I hope its popularity and usefulness grows. I'm truly looking forward to more people getting involved. The curren

Re: [convert] API Discussion...

2011-11-04 Thread James Carman
A branch would work just fine for that situation. Also, let's keep in mind that this component is in the sandbox On Nov 4, 2011 7:28 AM, "Adrian Crum" wrote: > Not so that someone else can commit them, so that others can review them > and comment on them. > > -Adrian > > On 11/4/2011 11:25 AM, Ja

Re: [convert] API Discussion...

2011-11-04 Thread Adrian Crum
Not so that someone else can commit them, so that others can review them and comment on them. -Adrian On 11/4/2011 11:25 AM, James Carman wrote: If need be, I would just create a branch for my work. It would be silly for me to submit patches so that someone else would commit them On Nov 4, 20

Re: [convert] API Discussion...

2011-11-04 Thread James Carman
Sounds interesting. I will try to take a deeper dive look at it this weekend On Nov 4, 2011 7:19 AM, "Adrian Crum" wrote: > It's a lot more dynamic than that. The factory is "smart" - it creates and > registers converters on-the-fly as it is used. > > -Adrian > > On 11/4/2011 11:17 AM, James Carm

Re: [convert] API Discussion...

2011-11-04 Thread James Carman
If need be, I would just create a branch for my work. It would be silly for me to submit patches so that someone else would commit them On Nov 4, 2011 7:20 AM, "Adrian Crum" wrote: > From my perspective, it would be preferable to keep the community involved > in the design decisions. > > -Adrian

Re: [convert] API Discussion...

2011-11-04 Thread Adrian Crum
From my perspective, it would be preferable to keep the community involved in the design decisions. -Adrian On 11/4/2011 11:15 AM, James Carman wrote: I don't have to submit a patch. I am a commons committer On Nov 4, 2011 5:55 AM, "Adrian Crum" wrote: The source and target classes are used

Re: [convert] API Discussion...

2011-11-04 Thread Adrian Crum
It's a lot more dynamic than that. The factory is "smart" - it creates and registers converters on-the-fly as it is used. -Adrian On 11/4/2011 11:17 AM, James Carman wrote: I would think that the source and target would be a part of the registration process and not a member of each individual

Re: [convert] API Discussion...

2011-11-04 Thread James Carman
I would think that the source and target would be a part of the registration process and not a member of each individual converter On Nov 4, 2011 5:55 AM, "Adrian Crum" wrote: > The source and target classes are used by the Converter.canConvert method. > The Converter.canConvert method is used by

Re: [convert] API Discussion...

2011-11-04 Thread James Carman
I don't have to submit a patch. I am a commons committer On Nov 4, 2011 5:55 AM, "Adrian Crum" wrote: > The source and target classes are used by the Converter.canConvert method. > The Converter.canConvert method is used by the Converter factory to find > the correct converter. The reason paramet

Re: [convert] API Discussion...

2011-11-04 Thread Adrian Crum
The source and target classes are used by the Converter.canConvert method. The Converter.canConvert method is used by the Converter factory to find the correct converter. The reason parameterized types are not used in this scenario is so you can create converters that handle entire class hierar

Re: [convert] API Discussion...

2011-11-04 Thread Simone Tripodi
Hi James, I had a look at current [convert] impl and the (source|target)Class are used to index converters inside a generic universal converter, so all you have to do is invoking just one method to perform conversions - without retrieving the converter you need on your client application. Your hint

[convert] API Discussion...

2011-11-03 Thread James Carman
I was taking a look at the [convert] component because I have done some work lately on some handy conversion classes. I'm struggling to understand why you'd need the getSourceClass() and getTargetClass() methods if you're using generics. Also, I've got a class that looks like this: public class