Devid HW
b843146554
adds lib target + cargo-fuzz fuzz targets exercising pure parsers; backs dynamic_analysis criteria and the Scorecard fuzzing check.
37 lines
1.3 KiB
Rust
37 lines
1.3 KiB
Rust
#![no_main]
|
|
|
|
use libfuzzer_sys::fuzz_target;
|
|
use mt5_quant::models::Metrics;
|
|
|
|
// Fuzz target for the pure, I/O-free MT5 backtest report metrics parser.
|
|
//
|
|
// `Metrics::from_html` takes an arbitrary `&str` and extracts the numeric
|
|
// summary metrics from an MT5 HTML/XML report. It is panic-free on bad input:
|
|
// every numeric parse uses `unwrap_or`, regex patterns are compile-time
|
|
// constants (never panicking), and the function returns `Option<Metrics>`
|
|
// (None when no metrics are found). We return early on non-UTF-8 input and on
|
|
// the None case, so the fuzzer only ever exercises the parser itself.
|
|
fuzz_target!(|data: &[u8]| {
|
|
let Ok(text) = std::str::from_utf8(data) else {
|
|
// Parser only accepts UTF-8 text; arbitrary byte sequences are
|
|
// out of scope and must not cause a crash.
|
|
return;
|
|
};
|
|
|
|
if let Some(metrics) = Metrics::from_html(text) {
|
|
// Force the computed fields to be materialized so any logic error
|
|
// inside the parser surfaces during fuzzing.
|
|
let _ = (
|
|
metrics.net_profit,
|
|
metrics.profit_factor,
|
|
metrics.max_dd_pct,
|
|
metrics.sharpe_ratio,
|
|
metrics.total_trades,
|
|
metrics.recovery_factor,
|
|
metrics.win_rate_pct,
|
|
metrics.gross_profit,
|
|
metrics.gross_loss,
|
|
);
|
|
}
|
|
});
|