[Feature]: reuse browserContext for debugging.

[ykawakamy]

🚀 Feature Request

I'd like to reuse the browser context during debugging.

Example

I propose test.sandbox annotation.

test("too much navigation", async({page})=>{
  await page.getByRole("button", { name:"page1"}).click();
  await page.getByRole("button", { name:"page2"}).click();
  await page.getByRole("button", { name:"page3"}).click();
  // ... too much navigation and too slow...

  // failed.
  await expect(page.locator("flaky-element")).toHaveText("xxx");
});


// proposal annotation
test.sandbox( "for fix flaky-element assertion", async ({page})=>{
  await expect(page.locator("flaky-element")).toHaveText("fixed");
});

when stabled,

 - await expect(page.locator("flaky-element")).toHaveText("xxx");
 + await expect(page.locator("flaky-element")).toHaveText("fixed");

...

-test.sandbox( "for fix flaky-element assertion", async ({page})=>{
-  await expect(page.locator("flaky-element")).toHaveText("fixed");
-});

test.sandbox use last browser context.

Motivation

Make tests for debugging to easy. (particularly, too slow navigation.)

--
sorry, i'm not good at English.

[mxschmitt]

This seems to be outside of the scope for Playwright, our core principle is to isolate tests from each-other. If you have to deal with slow navigation, we instead recommend test.step.