JDK-8308023 - Exponential classfile blowup with nested try/finally

[Archie Cobbs]

Hi Jan,

Thanks for the (helpful) comments.

On Fri, Jul 28, 2023 at 3:17 AM Jan Lahoda <jan.lahoda at oracle.com> wrote:

> Overall, given the usual complexity of doing anything in Gen, I believe
> the inclination was to see if we can wait until javac will use the
> ClassFile API to generate bytecode, and revisit then. This would at least
> help us generate proper StackMaps much more easier.
>

What's the timeline for this work?

Overall, I suspect that whatever we do, we would need to run proper
> experiments to ensure there's no problem with performance of the new code.
>

I agree that this will be an important litmus test for any new strategy.

But do you have a clear idea what "proper experiments" would look like? Are
there any simple off-the-shelf tests that would serve as reasonable initial
feedback while prototyping?

As they say, you can't improve (or tweak) what you don't measure, so it
seems a prerequisite is to properly define the "measure" and get agreement
on it.

I'm sure no one (including me) wants to invest in experimentation without
some established goalposts that can't be moved after the fact by naysayers.

-Archie

-- 
Archie L. Cobbs
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.openjdk.org/pipermail/compiler-dev/attachments/20230728/bea201ca/attachment.htm>