fix(qe): support MySQL base64 newlines

[Weakky]

Overview

fixes prisma/prisma#23262

MySQL encodes base64 strings with a newline every 76 characters. Rust's base64 crate does not support newlines and suggests sanitizing the input before passing it down. This PR sanitizes MySQL's base64 outputs.

Credit to @aqrln for the fast™ sanitization.

With read loop

fn read_loop(s: &str, buf: &mut [u8]) {
    let mut pos = 0;

    for c in s.chars() {
        if pos == buf.len() {
            break;
        }

        if c != '\n' {
            buf[pos] = c as u8;
            pos += 1;
        }
    }
}

Current implementation

fn sanitize_base64<'a>(mut s: &str, buf: &'a mut [u8]) -> &'a [u8] {
    let mut pos = 0;

    while !s.is_empty() && pos < buf.len() {
        let nl = s.find('\n').unwrap_or(s.len());
        let len = nl.min(buf.len() - pos);
        buf[pos..pos + len].copy_from_slice(&s.as_bytes()[..len]);
        pos += len;
        s = &s[(nl + 1).min(s.len())..];
    }

    &buf[0..pos]
}

[github-actions]

WASM Query Engine file Size

Engine	This PR	Base branch	Diff
Postgres	2.056MiB	2.053MiB	2.986KiB
Postgres (gzip)	820.389KiB	820.249KiB	144.000B
Mysql	2.026MiB	2.022MiB	4.068KiB
Mysql (gzip)	808.375KiB	806.666KiB	1.709KiB
Sqlite	1.925MiB	1.922MiB	2.986KiB
Sqlite (gzip)	768.531KiB	768.553KiB	-23.000B

[codspeed-hq]

CodSpeed Performance Report

Merging #4955 will not alter performance

_{Comparing fix/mysql-base64 (62d56b4) with main (7484fd8)}

Summary

✅ 11 untouched benchmarks

[jkomyno]

Nice! For future reference, is there a link to the source of the benchmarks and plots above?

[Weakky]

Nice! For future reference, is there a link to the source of the benchmarks and plots above?

use criterion::{black_box, criterion_group, criterion_main, Criterion};

fn read_loop(s: &str, buf: &mut [u8]) {
    let mut pos = 0;

    for c in s.chars() {
        if pos == buf.len() {
            break;
        }

        if c != '\n' {
            buf[pos] = c as u8;
            pos += 1;
        }
    }
}

fn read_memcpy(mut s: &str, buf: &mut [u8]) {
    let mut pos = 0;

    while !s.is_empty() && pos < buf.len() {
        let nl = s.find('\n').unwrap_or(s.len());
        let len = nl.min(buf.len() - pos);
        buf[pos..pos + len].copy_from_slice(&s.as_bytes()[..len]);
        pos += len;
        s = &s[(nl + 1).min(s.len())..];
    }
}

fn prepare_string(len: usize) -> String {
    let mut s = String::with_capacity(len);

    for i in 0..len {
        s.push(if len % 77 == 0 {
            '\n'
        } else {
            (i % 26 + 97) as u8 as char
        })
    }

    s
}

fn benchmark(c: &mut Criterion) {
    let mut buf = [0; 1024];
    let s = prepare_string(buf.len());

    c.bench_function("read_loop", |b| {
        b.iter(|| {
            read_loop(black_box(&s), black_box(&mut buf));
        })
    });

    c.bench_function("read_memcpy", |b| {
        b.iter(|| {
            read_memcpy(black_box(&s), black_box(&mut buf));
        })
    });
}

criterion_group!(benches, benchmark);
criterion_main!(benches);