Re: [DrJava] BeanShell

[Brian S. <bs...@bs...>]

On BeanShell: The first prototype of an interpreter-based REPL for 
DrJava that I did was actually based on BeanShell. I switched to 
DynamicJava for reasons I don't remember at the moment -- it might have 
been language feature support they had at the time, perhaps lacking 
anonymous inner classes or something. Or it may have been an ease of 
integration issue, or even licensing at that time (I think they used to 
have a different license).

Anyway, BeanShell is certainly a more living project than DynamicJava 
and it sounds like it has improved some since I last checked it out. It 
sounds like not a bad idea to move to it, although the range of 
improvements sound like it wouldn't be worth your effort to spend too 
much time on it.

On the other hand, it shouldn't be too hard to implement. The REPL was 
designed from the beginning to allow swapping out the interpeter. Pretty 
much, you should be able to implement model.repl.JavaInterpreter and 
swap that in instead. (I think there is a bit of DynamicJava-specific 
code in other places, like handling the returned values/error messages, 
but it shouldn't be hard to clean up.)

One small concern to point out: Seems like they blew it with regard to 
returning the "no result" value, like DynamicJava did (before I fixed it 
in our subclass). From their docs:

If an evaluation of a statement or expression yields a "void" value; 
such as would be the case for something like a for-loop or a void type 
method invocation, eval() returns null.

This same problem in DynamicJava was the reason, if you left off the 
semicolon, you used to get "null" as the result for void-returning 
statements.

Anyhow, just wanted to throw in my few cents.

-brian