feat: add Python and Java regex engines

2026-07-27 02:02:21 +02:00
parent 4951c966d4
commit 7079cde15f
95 changed files with 8866 additions and 251 deletions
--- a/tests/browser/flavours.spec.ts
+++ b/tests/browser/flavours.spec.ts
@@ -0,0 +1,123 @@
+import { expect, test, type Page } from "@playwright/test";
+
+async function setEditor(page: Page, testId: string, value: string) {
+  await page.getByTestId(testId).locator(".cm-content").fill(value);
+}
+
+async function waitForReady(page: Page, timeout = 15_000) {
+  await expect(page.locator(".run-status")).toHaveClass(/status-ready/u, {
+    timeout,
+  });
+}
+
+test("CPython re executes and replaces with code-point-native offsets under the production CSP", async ({
+  page,
+}) => {
+  const pageErrors: string[] = [];
+  const consoleErrors: string[] = [];
+  page.on("pageerror", (error) => pageErrors.push(error.message));
+  page.on("console", (message) => {
+    if (message.type() === "error") consoleErrors.push(message.text());
+  });
+
+  const response = await page.goto("/deep/nested/regex/");
+  expect(response?.headers()["content-security-policy"]).toContain(
+    "script-src 'self' 'wasm-unsafe-eval'",
+  );
+  expect(response?.headers()["content-security-policy"]).not.toContain(
+    "'unsafe-eval'",
+  );
+  await waitForReady(page);
+  await page.getByLabel("Live").uncheck();
+  await page.getByRole("combobox", { name: "Flavour" }).selectOption("python");
+  await setEditor(page, "editor-regular-expression-pattern", "(?P<word>\\w+)");
+  await setEditor(page, "editor-test-text", "😀 café");
+  await page.getByRole("button", { name: "Run", exact: true }).click();
+  await waitForReady(page, 45_000);
+
+  await expect(page.locator(".run-status")).toContainText(
+    "native offsets code-point",
+  );
+  const match = page.locator(".extraction-row").first();
+  await expect(match).toContainText("UTF-16 range: 3…7");
+  await expect(match).toContainText("native range: 2…6 code-point");
+
+  await page.getByRole("button", { name: "Replace", exact: true }).click();
+  await setEditor(page, "editor-replacement-template", "[\\g<word>]");
+  await page.getByRole("button", { name: "Run", exact: true }).click();
+  await waitForReady(page, 45_000);
+  await expect(
+    page.getByTestId("editor-replacement-output").locator(".cm-content"),
+  ).toHaveText("😀 [café]");
+
+  await page.getByRole("button", { name: "Quick reference" }).click();
+  await expect(
+    page.getByRole("dialog", { name: "Quick reference" }),
+  ).toContainText("Python 3.14 re documentation");
+  await page.keyboard.press("Escape");
+  await page.getByRole("button", { name: "Capabilities" }).click();
+  const capabilities = page.getByRole("dialog", { name: "Capabilities" });
+  await expect(capabilities).toContainText("CPython re via Pyodide");
+  await expect(capabilities).toContainText("CPython 3.14.2 re");
+  expect(pageErrors).toEqual([]);
+  expect(consoleErrors).toEqual([]);
+});
+
+test("TeaVM java.util.regex executes Java-only syntax and numeric replacement with UTF-16 offsets", async ({
+  page,
+}) => {
+  const pageErrors: string[] = [];
+  page.on("pageerror", (error) => pageErrors.push(error.message));
+  await page.goto("/deep/nested/regex/");
+  await waitForReady(page);
+  await page.getByLabel("Live").uncheck();
+  await page.getByRole("combobox", { name: "Flavour" }).selectOption("java");
+  await setEditor(
+    page,
+    "editor-regular-expression-pattern",
+    "(?<word>[a-z]++)",
+  );
+  await setEditor(page, "editor-test-text", "😀 abc");
+  await page.getByRole("button", { name: "Run", exact: true }).click();
+  await waitForReady(page);
+
+  await expect(page.locator(".run-status")).toContainText(
+    "native offsets utf16",
+  );
+  const match = page.locator(".extraction-row").first();
+  await expect(match).toContainText("UTF-16 range: 3…6");
+  await expect(match).toContainText("native range: 3…6 utf16");
+
+  await page.getByRole("button", { name: "Replace", exact: true }).click();
+  await setEditor(page, "editor-replacement-template", "<$1>");
+  await page.getByRole("button", { name: "Run", exact: true }).click();
+  await waitForReady(page);
+  await expect(
+    page.getByTestId("editor-replacement-output").locator(".cm-content"),
+  ).toHaveText("😀 <abc>");
+
+  await page.getByRole("button", { name: "Capabilities" }).click();
+  const capabilities = page.getByRole("dialog", { name: "Capabilities" });
+  await expect(capabilities).toContainText("TeaVM java.util.regex");
+  await expect(capabilities).toContainText("not OpenJDK");
+  expect(pageErrors).toEqual([]);
+});
+
+test("serves the self-hosted Python and Java runtime assets with executable MIME types", async ({
+  request,
+}) => {
+  const pythonModule = await request.get("./engines/python/pyodide.asm.mjs");
+  expect(pythonModule.ok()).toBe(true);
+  expect(pythonModule.headers()["content-type"]).toContain("text/javascript");
+
+  const pythonWasm = await request.get("./engines/python/pyodide.asm.wasm");
+  expect(pythonWasm.ok()).toBe(true);
+  expect(pythonWasm.headers()["content-type"]).toContain("application/wasm");
+  expect((await pythonWasm.body()).subarray(0, 4)).toEqual(
+    Buffer.from([0x00, 0x61, 0x73, 0x6d]),
+  );
+
+  const javaModule = await request.get("./engines/java/java-regex.mjs");
+  expect(javaModule.ok()).toBe(true);
+  expect(javaModule.headers()["content-type"]).toContain("text/javascript");
+});