Any plans to support type-safe access to java.lang.Method?

[Brian Goetz]

He's talking about something along the lines of the Coin proposal 
"Method and Field literals".  (The # notation is taken from the strawman 
syntax for lambdas in Java SE 8.)  The idea is that there'd be some 
{Method,Field}-valued expression like "ClassName#methodName" where the 
compiler verified the method was there and then emitted the appropriate 
reflective goo (catching reflective errors and wrapping them as runtime 
exceptions, since you now have some reason to believe the symbol 
actually exists (that's the type-safe part.)  It would let you write 
APIs that abstract more easily over member references.

The answer is: no, we're not considering this at this time.



On 6/19/2011 4:53 PM, Erik Trimble wrote:
> On 6/19/2011 5:07 AM, Behrang Saeedzadeh wrote:
>> Hi all,
>>
>> Are there any plans to support type-safe access to methods?
>>
>> For example something like:
>>
>> java.lang.Method toString = Object#toString();
>>
>> It would make writing APIs like JPA 2's type-safe criteria builder
>> more concise and readable.
>>
>>
>> Cheers,
>> Behrang Saeedzadeh
>> http://www.behrang.org
>
> Behrang,
>
> Exactly what do you mean? While Java (AFAIK) isn't a provably type-safe
> language (as such formal proofs are very difficult for most complex
> languages), it already has both runtime and compile time type checking,
> and does fit the practical description of a type-safe language. That is,
> Java already has type-safe access to methods.
>
> http://www.securingjava.com/chapter-two/chapter-two-10.html
>
> Also, the '#' sign isn't any sort of Java operator. I'm assuming that it
> comes from Perl (I though maybe C#, but it doesn't have it either), but,
> given the syntax, that doesn't make sense either. I don't understand the
> right side of your example, as I don't know what it's trying to accomplish.
>
> Also, since Java is a polymorphic language, you wouldn't want *too*
> strong a type-checking system, as I would always want the ability for a
> method to work on a subclass of any parameter I had defined the method
> as accepting.
>
>