Andrea Peruffo

Results 396 comments of Andrea Peruffo

Hi @kateinoigakukun and thanks for the great work! I'd like to provide some additional data points, from [this brief](https://github.com/dylibso/chicory/pull/972) exploration: - an hello world program currently compiles to a ~6.8M...

@ehsankianifar great work narrowing down the issue! > This issue was originally reported by ... Appreciated, but no need to carry around attributions πŸ™ > OpenJDK project I never reported...

Thanks for the amazing project and sorry for bumping such an old issue. In [Chicory](https://github.com/dylibso/chicory) we do [hit the JVM limits](https://github.com/dylibso/chicory/issues/405) for a single method size. Having a way to...

Thanks a lot @kripken for getting back! > Is there no JVM flag to avoid the issue for you? Unfortunately not, this is a hard limit very deeply encoded. >...

Thanks again for the feedback! > Does `wasm-opt -O3` not help here? It does in most cases(e.g. with `CPython`), I'll update this issue as soon as I have a new...

Ok, found 10 minutes to provide one example reproducer with Prism: - given this build artifact: https://github.com/ruby/prism/actions/runs/13090799608/artifacts/2521567203 - the function `pm_dump_json` is producing a huge single wasm function - even...

Found an easily inspectable Go package! - from this archive: https://registry.npmjs.org/esbuild-wasm/-/esbuild-wasm-0.25.0.tgz - extract `package/esbuild.wasm` - function 3093 is the first hitting the limit In this case `wasm-opt` takes a super...

Ping, is there any update on the subject? πŸ™‚ πŸ™ I'm especially interested if you have specific suggestions(configuration options?) to split the big functions [of SQLite](https://github.com/roastedroot/sqlite4j/tree/main/wasm-lib). Apparently `wasi-sdk` is already...

Thanks for getting back @tlively and thanks for the guidance. I'll keep this on my watch list, if no one comes first to it I might be able to dedicate...

> Β it is unrealistic that we can divide everything strictly by proposal Right, and we have mechanisms in place to exclude even single assertions, but having cross cutting concerns, for...