PROPOSAL: Compiletime information access

[rssh at gradsoft.com.ua]

> rssh at gradsoft.com.ua wrote:
> [snip]
>>>>  -  improving proposal about access to compile-time information about
>>>> local context can have some sence (?)
>>>>
>>>>
>>> Given the existing capabilities of annotation processors, I don't think
>>> the proposed API is necessary or desirable.
>>>
>> </cite>
>> <IMHO>
>>  I does not understand, why Joseph Darcy think, that existing
>> capabilities
>> of annotation processing may allows programmer use compile-time
>> information about local context (it's not true for now).
>> </IMHO>
>>
>
> Joe Darcy thinks standardized annotation processing is an existing
> feature in the platform to address problems in the space of analyzing
> programs and generating code.  The first-line solution to problems in
> this space is to approach the problem in a way that can be addressed by
> the existing annotation processing facilities.  Even if a particular
> worthwhile issue in this class of problem cannot be solving by
> annotation processing (or a small extension to annotation processing),
> we will probably choose to not provide a solution to the problem in the
> platform because the marginal utility of the solution over annotation
> processing is likely to be too small.
>

I. e. right place for such changes is annotations API. (And in general
 library-level changes is preferable agains language changes with same
effects).

Thanks for refinement.



> -Joe
>