Review of JEP draft

[David Holmes]

On 19/03/2022 7:16 pm, Julian Waters wrote:
> Hi David,
> 
> I was under the impression that the reason the standard for 
> introducing new opcodes is so stringent was partially due to the 
> constraints imposed by the current single byte scheme. Is that not the 
> case in practice? 

I think even without limited bytecodes the costs of adding new bytecodes 
are high enough for it to only be considered in very worthy cases.

I was trying to find a good, old, article that talks about the pros and 
cons of growing the bytecode but alas I could not. But some 
presentations from John Rose do touch on some aspects of this:

http://cr.openjdk.java.net/~jrose/pres/

Cheers,
David

> In any case, I'll keep this JEP shelved in case it may 
> be of use in the future.
> 
> best regards,
> Julian
> 
> On Fri, Mar 18, 2022 at 8:30 AM David Holmes <david.holmes at oracle.com 
> <mailto:david.holmes at oracle.com>> wrote:
> 
>     Hi Julian,
> 
>     On 17/03/2022 10:19 pm, Julian Waters wrote:
>      > Hi everyone,
>      >
>      > If you don't mind a little reading, can I get a review of the
>     following JEP
>      > draft at
>      >
>     https://bugs.openjdk.java.net/projects/JDK/issues/JDK-8283291?filter=allissues&orderby=created+DESC%2C+priority+DESC%2C+updated+DESC
>     <https://bugs.openjdk.java.net/projects/JDK/issues/JDK-8283291?filter=allissues&orderby=created+DESC%2C+priority+DESC%2C+updated+DESC>?
>      > Apologies if this is not the proper way to submit a JEP, I'm a
>     little new
>      > to this.
> 
>     The bar is set very, very high, for introducing new bytecodes and
>     consequently running out of them has not been a problem in practice.
>     The
>     reason the bar has been set so high is because the impact of a new
>     bytecode on the whole Java ecosystem is enormous. Numerous new features
>     have considered the possibility of adding a new bytecode, but very few
>     have actually done so, instead flexible mechanisms like invokeDynamic,
>     were introduced, that could then be used to implement a range of other
>     features.
> 
>     If we had almost no spare bytecodes left, and we regularly added new
>     bytecodes, then this would be a problem that needs solving. But as it
>     stands I don't see a real problem that needs solving here.
> 
>     YMMV.
> 
>     Cheers,
>     David
> 
>      > Have a great day!
>      >
>      > best regards,
>      > Julian
>