Adam Heath wrote:
> Adrian Crum wrote:
>> --- On Tue, 2/9/10, Adam Heath <
[hidden email]> wrote:
>>
>>> From: Adam Heath <
[hidden email]>
>>> Subject: conversion framework oddity
>>> To:
[hidden email]
>>> Date: Tuesday, February 9, 2010, 1:12 AM
>>> I'm trying to convert a String to a
>>> String, using the new conversion
>>> system written by Adrian. This fails, due to it's use
>>> of a static
>>> NullConvertor, that always returns Object for the source
>>> and target
>>> classes.
>>>
>>> I see that this NullConvertor is used to provide for
>>> pass-thru style
>>> conversions, but, in doing so, breaks the contract of the
>>> conversion
>>> system, by not returning the proper type for source and
>>> target.
>>>
>>> Adrian, do you agree with this? If so, I will fix
>>> it.
>> I'm sure whatever you have in mind will be fine. I was trying to keep things simple, but you're right about not returning the proper types.
>
> I've got a fix staged.
>
> Btw, I had something very similiar to this inside webslinger; however,
> it only did conversion to/from string, not the entire grid of
> possibilities. I've actually been able to use this library directly,
> and in a way you never thought. This is a good thing.
>
> As another bonus, the test cases I already had for my existing
> conversion code are working with this new library.
That's good to know. As you know, the conversion framework started out
being nothing more than an effort to replace a long switch statement
with polymorphism.
Btw, the integration of the conversion framework into the entity engine
is not complete. There is a patch that needs to be tested against
various databases:
https://issues.apache.org/jira/browse/OFBIZ-3245-Adrian
Locked
