jastrachan | 7 Oct 10:55 2004

Re: [groovy-dev] making the bytecode generation more understandable

On 7 Oct 2004, at 09:48, John Wilson wrote:
> On 7 Oct 2004, at 09:29, John Rose wrote:
>> On balance, I feel it is "beyond groovy" (dare I say, "techno punk"?) 
>> to have a simple identifier perform multiple runtime lookups to find 
>> its meaning.  (There's also the scary throw/catch backtracking in 
>> Closure.java.)
> The scary throw/catch backtracking in Closure (which is repeated for 
> normal method invocation) is caused by a poor design choice in the 
> InvokerHelper (no offence meant  Bob and/or James, I understand that 
> it was important to get the stuff working quickly). It introduces a 
> big performance hit, several bugs and is one of the main contributors 
> to debugging hell. It is, however, an implementation issue. It appears 
> to me that it would be perfectly possible to produce a simple dispatch 
> mechanism which does not (ab)use the exception mechanism,

I await your patch/fixes with excited anticipation :)