Files
regex-tools/tests/conformance/rust.conformance.test.ts

279 lines
8.4 KiB
TypeScript

// @vitest-environment node
import { createHash } from "node:crypto";
import { readFile } from "node:fs/promises";
import path from "node:path";
import { pathToFileURL } from "node:url";
import { afterAll, beforeAll, describe, expect, it } from "vitest";
import {
RUST_ENGINE_VERSION,
RUST_FLAVOUR_VERSION,
RustEngineAdapter,
} from "../../src/regex/execution/adapters/rust/RustEngineAdapter";
import {
RustWasmBridge,
type RustRegexWasmModule,
} from "../../src/regex/execution/adapters/rust/RustWasmBridge";
import type {
RegexExecutionRequest,
RegexReplacementRequest,
} from "../../src/regex/model/match";
const pack = path.resolve("public/engines/rust");
let adapter: RustEngineAdapter;
function request(
pattern: string,
subject: string,
overrides: Partial<RegexExecutionRequest> = {},
): RegexExecutionRequest {
return {
flavour: "rust",
flavourVersion: RUST_FLAVOUR_VERSION,
pattern,
flags: ["g"],
subject,
captureMetadata: [],
scanAll: false,
maximumMatches: 100,
maximumCaptureRows: 1_000,
...overrides,
};
}
function replacement(
pattern: string,
subject: string,
replacementText: string,
overrides: Partial<RegexReplacementRequest> = {},
): RegexReplacementRequest {
return {
...request(pattern, subject),
replacement: replacementText,
maximumOutputBytes: 1_024,
...overrides,
};
}
beforeAll(async () => {
const imported = (await import(
`${pathToFileURL(path.join(pack, "rust-regex.mjs")).href}?conformance`
)) as RustRegexWasmModule;
imported.initSync({
module: await WebAssembly.compile(
await readFile(path.join(pack, "rust-regex_bg.wasm")),
),
});
adapter = new RustEngineAdapter(
new RustWasmBridge(imported),
"Node.js conformance",
);
await adapter.load();
});
afterAll(() => {
adapter?.terminate();
});
describe("Rust regex 1.13.1 WebAssembly conformance", () => {
it("loads the exact crate and compiler identity", async () => {
await expect(adapter.load()).resolves.toEqual(
expect.objectContaining({
flavour: "rust",
engineName: "Rust regex crate",
engineVersion: RUST_ENGINE_VERSION,
runtimeVersion: expect.stringContaining(
"rustc 1.97.1 wasm32-unknown-unknown",
),
offsetUnit: "utf8-byte",
}),
);
});
it("returns native UTF-8 byte offsets and named captures", async () => {
const result = await adapter.execute(
request("(?P<emoji>😀+)([A-Z]+)?", "x😀😀AB 😀"),
);
expect(result.accepted).toBe(true);
expect(
result.matches.map(({ nativeRange, range }) => ({ nativeRange, range })),
).toEqual([
{
nativeRange: { start: 1, end: 11, unit: "utf8-byte" },
range: { startUtf16: 1, endUtf16: 7 },
},
{
nativeRange: { start: 12, end: 16, unit: "utf8-byte" },
range: { startUtf16: 8, endUtf16: 10 },
},
]);
expect(result.matches[0]?.captures[0]).toEqual(
expect.objectContaining({
groupNumber: 1,
groupName: "emoji",
value: "😀😀",
nativeRange: { start: 1, end: 9, unit: "utf8-byte" },
}),
);
expect(result.matches[1]?.captures[1]).toEqual({
groupNumber: 2,
status: "did-not-participate",
});
});
it("uses Rust's native Unicode default and rejects unsupported constructs", async () => {
const unicode = await adapter.execute(
request("^\\w+$", "Grüße", { flags: [] }),
);
expect(unicode.matches).toHaveLength(1);
const unsupported = await adapter.execute(request("(?<=a)b", "ab"));
expect(unsupported.accepted).toBe(false);
expect(unsupported.diagnostics[0]).toEqual(
expect.objectContaining({ code: "compile-error" }),
);
});
it("maps i, m, s, U, u, x and R through RegexBuilder", async () => {
const cases = [
["a", "A", ["i"]],
["^b$", "a\nb\nc", ["m"]],
["a.b", "a\nb", ["s"]],
["a+?", "aaa", ["U"]],
["^ a + $", "aaa", ["x"]],
["^b$", "a\r\nb\r\nc", ["m", "R"]],
["\\p{Greek}+", "λ", ["u"]],
] as const;
for (const [pattern, subject, flags] of cases) {
const result = await adapter.execute(
request(pattern, subject, { flags: [...flags] }),
);
expect(result.matches, `${flags.join("")}: ${pattern}`).toHaveLength(1);
}
const ungreedy = await adapter.execute(
request("a+", "aaa", { flags: ["U"] }),
);
expect(ungreedy.matches[0]?.value).toBe("a");
});
it("uses native Captures::expand replacement syntax", async () => {
const result = await adapter.replace(
replacement(
"(?P<word>[a-z]+)-(\\d+)",
"a-1 b-22",
"${word}[$2]/$word/$$",
),
);
expect(result.execution.accepted).toBe(true);
expect(result.output).toBe("a[1]/a/$ b[22]/b/$");
const longest = await adapter.replace(replacement("(a)", "a", "$1x|${1}x"));
expect(longest.output).toBe("|ax");
});
it("bounds capture amplification and preserves the untouched partial suffix", async () => {
const bounded = await adapter.replace(
replacement("(a+)", "a".repeat(256 * 1024), "$1".repeat(4_096), {
maximumOutputBytes: 5,
}),
);
expect(bounded.execution.accepted).toBe(true);
expect(bounded.output).toBe("aaaaa");
expect(bounded.outputBytes).toBe(5);
expect(bounded.outputTruncated).toBe(true);
const partial = await adapter.replace(
replacement("a", "aaa", "x", { maximumMatches: 1 }),
);
expect(partial.execution.matches).toHaveLength(1);
expect(partial.execution.truncated).toBe(true);
expect(partial.output).toBe("xaa");
expect(partial.outputTruncated).toBe(false);
});
it("enforces match/capture/output bounds without splitting UTF-8", async () => {
const limited = await adapter.execute(
request("a", "aaa", { maximumMatches: 1 }),
);
expect(limited.matches).toHaveLength(1);
expect(limited.truncated).toBe(true);
const rows = await adapter.execute(
request("(a)(b)", "ab", { maximumCaptureRows: 1 }),
);
expect(rows.matches).toHaveLength(0);
expect(rows.truncated).toBe(true);
const output = await adapter.replace(
replacement(".", "abc", "😀", { maximumOutputBytes: 5 }),
);
expect(output.output).toBe("😀");
expect(output.outputBytes).toBe(4);
expect(output.outputTruncated).toBe(true);
});
it("rejects lone UTF-16 surrogates before entering Rust", async () => {
await expect(adapter.execute(request(".", "\ud800"))).rejects.toThrow(
/unpaired UTF-16 surrogate/u,
);
});
it("accepts control-heavy subjects within the decoded 16 MiB contract", async () => {
const subject = "\u0000".repeat(3 * 1024 * 1024);
const result = await adapter.execute(
request("^\\x00+$", subject, {
flags: [],
maximumMatches: 1,
maximumCaptureRows: 1,
}),
);
expect(result.accepted).toBe(true);
expect(result.matches[0]?.nativeRange).toEqual({
start: 0,
end: subject.length,
unit: "utf8-byte",
});
});
it("ships verified metadata, licences, Cargo.lock and checksums", async () => {
const metadata = JSON.parse(
await readFile(path.join(pack, "engine-metadata.json"), "utf8"),
) as {
engineVersion: string;
source: { crateChecksumSha256: string; license: string };
toolchain: {
rustcCommitSha1: string;
wasmBindgenVersion: string;
};
};
expect(metadata).toEqual(
expect.objectContaining({
engineVersion: "1.13.1",
source: expect.objectContaining({
crateChecksumSha256:
"f020237b6c8eed93db2e2cb53c00c60a8e1bc73da7d073199a1180401450218d",
license: "MIT OR Apache-2.0",
}),
toolchain: expect.objectContaining({
rustcCommitSha1: "8bab26f4f68e0e26f0bb7960be334d5b520ea452",
wasmBindgenVersion: "0.2.126",
}),
}),
);
const checksums = await readFile(path.join(pack, "SHA256SUMS"), "utf8");
for (const line of checksums.trim().split("\n")) {
const [expected, name] = line.split(/ {2}/u);
const value = await readFile(path.join(pack, name ?? ""));
expect(createHash("sha256").update(value).digest("hex"), name).toBe(
expected,
);
}
expect(
await readFile(path.join(pack, "LICENSE-MIT.txt"), "utf8"),
).toContain("Permission is hereby granted");
expect(
await readFile(path.join(pack, "LICENSE-Apache-2.0.txt"), "utf8"),
).toContain("Apache License");
});
});