Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Use new Roast library for custom AST serialization #1001

Merged
merged 1 commit into from
Aug 21, 2024
Merged

Conversation

anderseknert
Copy link
Member

And the jsoniter library for any JSON handling.

This is quite a big win for linting performance:

➜ hyperfine --warmup 2 'regal_main lint bundle' 'regal_jsoniter lint bundle'
Benchmark 1: regal_main lint bundle
  Time (mean ± σ):      2.707 s ±  0.035 s    [User: 19.171 s, System: 0.636 s]
  Range (min … max):    2.662 s …  2.785 s    10 runs

Benchmark 2: regal_jsoniter lint bundle
  Time (mean ± σ):      2.185 s ±  0.019 s    [User: 14.932 s, System: 0.466 s]
  Range (min … max):    2.147 s …  2.207 s    10 runs

Summary
  regal_jsoniter lint bundle ran
    1.24 ± 0.02 times faster than regal_main lint bundle

But importantly, it doesn't make the AST harder to read or work with — quite the opposite actually!

See the Roast README for all the details:
https://github.com/anderseknert/roast

And the `jsoniter` library for any JSON handling.

This is quite a big win for linting performance:

```
➜ hyperfine --warmup 2 'regal_main lint bundle' 'regal_jsoniter lint bundle'
Benchmark 1: regal_main lint bundle
  Time (mean ± σ):      2.707 s ±  0.035 s    [User: 19.171 s, System: 0.636 s]
  Range (min … max):    2.662 s …  2.785 s    10 runs

Benchmark 2: regal_jsoniter lint bundle
  Time (mean ± σ):      2.185 s ±  0.019 s    [User: 14.932 s, System: 0.466 s]
  Range (min … max):    2.147 s …  2.207 s    10 runs

Summary
  regal_jsoniter lint bundle ran
    1.24 ± 0.02 times faster than regal_main lint bundle
```

But importantly, it doesn't make the AST harder to read or work with — quite
the opposite actually!

See the Roast README for all the details:
https://github.com/anderseknert/roast

Signed-off-by: Anders Eknert <anders@styra.com>
Copy link
Member

@srenatus srenatus left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM. It's a bit of a bummer that we can't measure the impact of jsoniter alone vs perf improvements from the AST change. 🤔

@anderseknert anderseknert merged commit f28ac7d into main Aug 21, 2024
3 checks passed
@anderseknert anderseknert deleted the roast branch August 21, 2024 08:14
@anderseknert
Copy link
Member Author

@srenatus yeah ... we'd have to use jsoniter in OPA for that too, or else all the marshalling of AST nodes will just pass through the encoding/json implementations. For the few places where we're serializing other things, like capabilities, the difference was negligible.

srenatus pushed a commit to srenatus/regal that referenced this pull request Oct 1, 2024
And the `jsoniter` library for any JSON handling.

This is quite a big win for linting performance:

```
➜ hyperfine --warmup 2 'regal_main lint bundle' 'regal_jsoniter lint bundle'
Benchmark 1: regal_main lint bundle
  Time (mean ± σ):      2.707 s ±  0.035 s    [User: 19.171 s, System: 0.636 s]
  Range (min … max):    2.662 s …  2.785 s    10 runs

Benchmark 2: regal_jsoniter lint bundle
  Time (mean ± σ):      2.185 s ±  0.019 s    [User: 14.932 s, System: 0.466 s]
  Range (min … max):    2.147 s …  2.207 s    10 runs

Summary
  regal_jsoniter lint bundle ran
    1.24 ± 0.02 times faster than regal_main lint bundle
```

But importantly, it doesn't make the AST harder to read or work with — quite
the opposite actually!

See the Roast README for all the details:
https://github.com/anderseknert/roast

Signed-off-by: Anders Eknert <anders@styra.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants