Ancestors

Toot

Written by Ed Page on 2025-01-06 at 20:56

Anyone have experience looking at #rustlang self-profile's and want to help interpret where Winnow's build times are going? My original hypothesis (return position impl trait) doesn't seem to be panning out and I'm not making sense of the profiling data to form a new one though maybe there is some rustc context I don't have.

Repo: https://github.com/winnow-rs/winnow

Benchmark command: cargo build --example json

=> More informations about this toot | More toots from epage@hachyderm.io

Descendants

Written by Manpacket on 2025-01-06 at 22:10

@epage clean release build of winnow itself + example takes 0.35s, clean release build for all the things is 6.34s, neither time seem unreasonable to me. What time do you see/expecting to see?

=> More informations about this toot | More toots from manpacket@functional.cafe

Written by Ed Page on 2025-01-06 at 22:50

@manpacket On https://github.com/rosetta-rs/parse-rosetta-rs?tab=readme-ov-file#results I get 2s for a debug build of an application on my benchmarking machine (its slow)