Divy Srivastava

The op overhead is most likely the cause. Started since #10844 which moved decoding to Rust. See below, ops dispatched when decoding was done in JS v/s in Rust: ```...

@vslinko I've put together a PR to optimize for this. https://github.com/denoland/deno/pull/14895 - after applying the patch, I get similar results as Deno 1.10.3 by tweaking chunk size to 1024 bytes....

Actionable items: - [x] Flatten op arguments. Deserializing `serde_v8` struct keys is 18% overhead! https://github.com/denoland/deno/pull/15723 - [ ] `TypedArrayPrototypeSlice` is 17% overhead, why do we need to do that? It...

hi @kriszyp 👋 thanks for looking into it I'm getting similar error on macOS with lmdb 2.3.0-alpha.1. It seems fail to instantly for me with `dyld[9373]: missing symbol called`. This...

@kriszyp > `thread 'main' has overflowed its stack` ok, It seems to be coming from the test harness powering `cargo test` and not the actual Deno subprocess. I might be...

> I think we can start by releasing a version based on the current implementation. I see that the APIs that have been implemented so far are sufficient to support...

Thank you @DjDeveloperr and @bartlomieju for the help!

It's not TextEncoder, `encode` function returns back the input its a TypedArray

For the record, here's the link to the conversation on V8 supporting strings in Fast API calls. https://discord.com/channels/684898665143206084/778060818046386176/1008754326933618728

After applying the encodeInto optimization in #15922