[Feature]: support synchronous test.step about playwright

[pavelfeldman]

It is unlikely that we make steps for sync operations since e2e tests are inherently async and there is an await workaround. Closing as out of scope for Playwright.

Btw, this.page.locator('.task-item').filter({ has: / * some other complex locators */}) isn't really "getting" the locator, it just builds a selector string in memory for this locator.

[alvaromartmart]

It is unlikely that we make steps for sync operations since e2e tests are inherently async and there is an await workaround. Closing as out of scope for Playwright.

Fair enough. I was actually wondering, while writing the issue, if a synchronous test.step could actually break functionality in reporter(s), for instance. It was a QoL ask anyway.

Btw, this.page.locator('.task-item').filter({ has: / * some other complex locators */}) isn't really "getting" the locator, it just builds a selector string in memory for this locator.

I understand how Playwright locators work 😄 - I don't see what's wrong with my statement or example code, the method does return a Locator. It just encapsulates the logic inside a POM method so the person developing the test doesn't need to remember how to build that specific locator.

This is what I would normally do, with the -supported- async test.step

// POM
class TodoPage {
    // [...]
   getLocatorForTask(contents: string): Promise<Locator> {
     return test.step(`[TodoPage] get locator for task containing "${contents}" in either name or description`, () => {
       return this.page.locator('.task-item').filter({ has: / * some other complex locators */});
     });
   }
}
// test
await expect(await todoPage.getLocatorForTask('foo')).toBeVisible();

(I can of course get rid of the cosmetic test.step and make the method syncrhonous, but that wasn't the main point - maybe I didn't pick the best example)