Re: Why isn't ConstantPoolGen final?

You never know what people did out there. No clue. Once it's open -  
it's open.

The problem in changing it will mainly be backwards compatibility.

cheers
--
Torsten

On 05.02.2008, at 14:13, Markus Gaisbauer wrote:

> Hi!
>
> After my recent performance patch, I have further looked for code  
> that could possibly be speed by applying a few optimisations.
>
> Although only a small percentage of total time is spend in ClassGen  
> initialisation (about 20% of deserialization and about 10% of  
> combined deserialization/serialization time), some speedup may be  
> achieved here.
>
> I was wondering why ConstantPoolGen isn't final, and why the fields  
> size, constants and index as well as the method adjustSize are  
> protected? Are there actually subclasses out there that depend on  
> their contracts implicitly defined by the current implementation?
>
> cheers,
> Markus
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [email protected]
> For additional commands, e-mail: [email protected]
>