Tomasz Andrzejak

Let me know what else I can check. I'm happy to dig into this :)

OS: Pop!_OS 22.04 LTS Linux version: 6.0.6-76060006-generic ([email protected]) It is based on Ubuntu 22.04 LTS, but they provide updated kernel, I guess.

Sure! Here is the diff at about 10k frames collected over all visible frames (~130 frames) Before:  After: 

Bench suite results: ``` Running benches/benchmark.rs (target/release/deps/benchmark-901adb2054ed4cfe) Gnuplot not found, using plotters backend profile_function time: [94.230 ns 94.448 ns 94.693 ns] change: [-2.6028% -2.2195% -1.8438%] (p = 0.00 < 0.05)...

The only relevant benchmark for the changes is `flush_frames`. We could expect `flush_frames` performance to regress due to added complexity, specifically caching stats on `add_frame` calls but the benches show...

@kondrak I will add an entry to changelog as requested by @Hoodad and it should be ready to go. Edit: Done :heavy_check_mark:

> that this is now blocked on the Spidermonkey upgrade path for StarlingMonkey Is this about moving away from forked SpiderMonkey and inlining the streams implementation in StarlingMonkey?

Hey @tschneidereit would you mind if I try to revive this PR? :)

An example test failure: ``` Serving HTTP on http://0.0.0.0:36005/ stderr [0] :: Warning: Using a DEBUG build. Expect things to be SLOW. stderr [0] :: [-1] Assertion failure: end >=...

I’m going to start working on this now.