diff options
| author | Matthias Krüger <matthias.krueger@famsik.de> | 2022-02-19 06:45:30 +0100 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2022-02-19 06:45:30 +0100 |
| commit | 554aea90b8e382678de0e87e07210af19f67f0ab (patch) | |
| tree | 715280de5ddbab0e6fb3947b29838badafe6ac7b /src/test/rustdoc-js-std/parser-errors.js | |
| parent | f19adc7acc649ad2b18b6e172b683ebadb3d8a92 (diff) | |
| parent | ae158224acb8a85efce18cd358a445d18c6c6389 (diff) | |
| download | rust-554aea90b8e382678de0e87e07210af19f67f0ab.tar.gz rust-554aea90b8e382678de0e87e07210af19f67f0ab.zip | |
Rollup merge of #93954 - aDotInTheVoid:json-buffer, r=Mark-Simulacrum
rustdoc-json: buffer output
It turns out we were doing syscalls for each part of the json syntax
Before:
```
...
[pid 1801267] write(5, "\"", 1) = 1
[pid 1801267] write(5, ",", 1) = 1
[pid 1801267] write(5, "\"", 1) = 1
...
```
After:
```
[pid 1974821] write(5, "{\"root\":\"0:0\",\"crate_version\":nu"..., 1575) = 1575
```
In one benchmark (one struct, almost all time in `std`), this gives ~2x perf
r? `@CraftSpider`
`@rustbot` modify labels: +A-rustdoc-json +T-rustdoc -A-testsuite
Diffstat (limited to 'src/test/rustdoc-js-std/parser-errors.js')
0 files changed, 0 insertions, 0 deletions