detailed_trace feature on wasm grows faster than on native
- Dominant language
- Rust
- Stars
- 48.2k
- Forks
- 4.8k
- Avg merge
- 3d 22h
- Merged PRs (30d)
- 161
Description
hello !
## Bevy version
0.16.0
## \[Optional\] Relevant system information
If you cannot get Bevy to build or run on your machine, please include:
cargo 1.87.0-nightly (307cbfda3 2025-03-20)
linux wayland
tested both latest firefox and chromium
## What you did
```
// can be any examples
cargo build --target wasm32-unknown-unknown --example sprite_sheet -F bevy/detailed_trace
wasm-server-runner target/wasm32-unknown-unknown/debug/examples/sprite_sheet.wasm
```
Observe the memory usage of the tab growing, for me its ~40mb/s or 1gb/25sec.
## What went wrong
the memory usage keeps growing until crash.
the performance of the bevy app is also greatly impacted, near 30% fps loss, with big drop to two digits fps every few seconds.
## Additional information
i dont think the performance loss is just the overhead of detailed trace. on mobile, i go from 144fps to 10. maybe just the allocation in loop.
this problem didn't occur on bevy 0.15.0 (the feature is new anyway i think?)
ive tried disabling logs at compile time with log and tracing features max level info but didn't work.
this might NOT be a memory leak, i just feel like its better for the title, as its more concrete and the end problem is the same.
Contributor guide
Assessment
This issue has not been assessed yet.