Categories
Playwright

Demonstrating the value of our Playwright tests

I don’t think anyone would disagree that creating and maintaining a set of e2e automated tests (like Playwright) takes a lot of time and effort, and therefore costs a lot of money.

So how does one demonstrate the value of investing time/effort/money into automated e2e testing?

My goal for automated e2e tests like Playwright is to have just enough that the team/project/product has zero manual regression testing required to release frequently.

I’ve written about this before, and it doesn’t mean no human testing, it just means no manual regression tests or test scripts a human follows to make sure existing functionality isn’t being broken by changes introduced into your system.

When looking at metrics I like the GQM approach where you start with goals, devise some questions to determine whether you’re achieving your goals, and create some metrics to answer those questions.

Our goal is already articulated above:

Goal: Zero manual regression testing

Some questions we could ask to see whether we’re achieving our goal would be:

  1. How good are our Playwright tests at catching regressions?
  2. How much manual regression testing do we perform?

And finally some metrics to answer our questions:

  1. Number of regression bugs found by the Playwright test suite
  2. Number of regression bugs not caught by the Playwright test suite
  3. Time spent performing manual regression testing

I like keeping things really simple when it comes to collecting and displaying metrics.

I created a Confluence page where I simply recorded regressions as they happened, in a table like such:

No.DateRegressionPlaywright/ManualRaised byStatus
1.10 JanWelcome screen doesn’t displayPlaywrightAlisterWIP
2.
Sample table to collect data

And using the table I created two simple graphs within the Confluence page to show our metrics:

Trend of regressions found
Playwright vs Manual

I think these metrics answer the question “how good are our Playwright tests at catching regressions?

To answer the other question “how much manual regression testing do we perform?” I can ask our QAs this question in our fortnightly catch up and record the results similar to the data above.

By using the answers to our questions we can determine whether or not we’re meeting our goal of having “zero manual regression testing” quite easily and whether Playwright is helping us achieve this goal.

What metrics do you collect around automated e2e testing?

Categories
Playwright

Debugging Playwright Tests with VS Code

I use VS Code as my text editor/IDE for writing Playwright tests. I can use use VS Code for debugging since it offers full debugging functionality like breakpoints and being able to see variables etc.

To enable this, there’s a couple of things you do:

  1. I created a debug task in my package.json file: "debug": "npx playwright test --headed --timeout=0" which means I can use npm run debug to execute a test without a timeout and showing the browser – by either adding a .only to a specific test, or telling it a file, eg. npm run debug ./scenarios/test1.spec.ts
  2. In VSCode use the “View → Debug Console” menu option, choose “Terminal” and make sure “JavaScript Debug Terminal” is set as the terminal type.
  3. Add a breakpoint in your code using the red dot in the left margin
  4. You can then use the npm run debug command which starts a debugging session where you can step through and see variables etc.
Debugging Playwright scripts in VS Code

Happy Debugging & Happy New Year! 🥳

Categories
e2e Testing Playwright

10 tips for successful e2e web app test automation

  1. Write independent automated tests: you should try remove dependencies on other tests or test data – this allows tests to be consistent, repeatable and to run in parallel (see #6).
  2. Set up data/state for each test via API calls: calling APIs is quick and efficient and can set up exactly what you need.
  3. Clean up data/state for each test using “after” hooks: this ensures test environments are kept clean and tidy and unwanted test data doesn’t cause issues with exploratory testing.
  4. Re-use browser authentication so you only need to log in once: this speeds up tests, see this post on how to do this with Playwright.
  5. Generate and use consistent (static) test data for each test: only generate unique/randomised values to satisfy uniqueness constraints, else use hard-coded known good values. Further reading here.
  6. Run all tests in parallel locally and in CI: hardware is powerful and there’s really no reason not to (unless you use Cypress and can’t 😝)
  7. Run new/updated tests at least 10 times locally in parallel before committing: this helps with reducing and removing non-deterministic tests and race conditions from your test suite.
  8. Use your automated test scripts to assist with manual/exploratory testing: for example you can easily set up state/accounts/sessions for testing – create npm commands so you can run npm run newuser for example to generate and log in as a brand new user ready for testing.
  9. Use linting/code autoformatting: such as JavaScript Standard Style, for consistently formatted code and not having to make decisions.
  10. Focus on reducing the need for manual regression testing, rather than code coverage: when you can confidently release your web application with no manual regression testing you know you have enough e2e automated tests.

What are your tips for successful web app test automation?

Categories
e2e Testing Playwright

Writing automated e2e tests for known buggy systems

Every system I’ve worked on, old or new, is full of known bugs (and unknown bugs for good measure 🤪). These known bugs are the ones that have never made it to the top of the bug backlog to be fixed because there’s always other more important work to do.

But what do you do with automated e2e tests that exercise such code and demonstrate such bugs?

Imagine a very simple example of a test that visits our page and asserts the title is correct.

Our page looks like this:

super simple webpage

And our Playwright code looks like this:

test.only('can have a test for a known bug in the system', async ({ page }) => {
  await goToPath(page, 'leave')
  expect(page.locator('#leavepage')).toHaveText('WebDriverJs Demo Leave Page');
})

You can see our test has different text it asserts than what is displayed. The text is our test is what we actually want to display, however the system displays it differently so our test fails when we run it.

What do we do with such tests? There’s a few different options all with their own advantages and disadvantages.

Option One: Commit the failing test as it is

Advantages: test is pure and correct, test is still run on every build highlighting the functionality that is wrong

Disadvantages: each build will fail until this functionality is fixed, creating red/failed builds and not giving immediate feedback on other potential issues found in the builds and resulting in people losing confidence in overall build results.

I personally wouldn’t recommend this approach as I think the noise of the failing builds outweighs any benefits it has.

Option Two: Mark the failing test as skipped

test.skip('can have a test for a known bug in the system', async ({ page }) => {
  await goToPath(page, 'leave')
  expect(page.locator('#leavepage')).toHaveText('WebDriverJs Demo Leave Page');
})

Advantages: no noise in builds since test no longer runs

Disadvantages: test can be forgotten about since it never runs and other issues could be introduced in the feature. For example if the text was changed to something else that is also wrong we wouldn’t know since the test is not being run.

Whilst this is preferable to option one this option often results in forgotten tests so I would also not recommend it.

Option Three: Update the assertion to be incorrect (with a comment)

test.only('can have a test for a known bug in the system', async ({ page }) => {
  await goToPath(page, 'leave')
  expect(page.locator('#leavepage')).toHaveText('WebDriverJs Demo Leaving Page'); // BUG: This text should be WebDriverJs Demo Leave Page
  })

Advantages: if the text changes to any value (whether now correct, or still incorrect) the test will fail alerting us to a change in functionality

Disadvantages: the tests are no longer representative of what is expected of the system – the expectations contradict what is actually expected.

I probably prefer this to having a pending test but something doesn’t feel right about a false assertion.

Option Four:

Playwright actually offers a solution for scenarios like this, it’s the test.fail() syntax which marks a test as being expected to fail, so it is still run but if it fails it passes, and if it passes it fails 🙃

We can write the test like this:

test.only('can have a test for a known bug in the system', async ({ page }) => {
  test.fail() // BUG: The text is presently wrong
  await goToPath(page, 'leave')
  expect(page.locator('#leavepage')).toHaveText('WebDriverJs Demo Leave Page');
  })

And when it fails it “passes” with a green cross:

If the system was fixed this test would then fail, and we’d know to remove the test.fail()line.

Advantages: if the text changes to the correct value we will know as this test will pass when we don’t expect it to. We can keep our assertions correct/pure.

Disadvantages: if the test was to fail in a different way we wouldn’t know about it since all the test cares about is that it fails (which we’re expecting).

Whilst this can hide other test failures, since I aim to write independent tests I can live with it potentially hiding other issues so this is my preferred approach to known failures.

How do you deal with known failures? Any of these ways or another I’ve missed?

Categories
Playwright

Reusable Authentication Across Playwright Tests

Most web apps require user authentication which requires the user to login at the start of an e2e test. A very basic example is:

import { test, expect } from '@playwright/test'

test.describe.parallel('Unauthenticated tests', () => {
  test('can view as guest', async ({ page }) => {
    await page.goto('http://webdriverjsdemo.github.io/auth/')
    await expect(page.locator('text=Hello Please Sign In')).toBeVisible()
  })
})

test.describe.parallel('Authenticated tests', () => {
  test('can view as admin', async ({ page }) => {
    await page.goto('http://webdriverjsdemo.github.io/auth/')
    await page.type('#firstname', 'Admin')
    await page.type('#surname', 'User')
    await page.click('#ok')
    await page.goto('http://webdriverjsdemo.github.io/auth/')
    await expect(page.locator('text=Welcome name=AdminUser')).toBeVisible()
  })

  test('can view as standard user', async ({ page }) => {
    await page.goto('http://webdriverjsdemo.github.io/auth/')
    await page.type('#firstname', 'Standard')
    await page.type('#surname', 'Person')
    await page.click('#ok')
    await page.goto('http://webdriverjsdemo.github.io/auth/')
    await expect(page.locator('text=Welcome name=StandardPerson')).toBeVisible()
  })
})

Whilst it’s easy to move the common code which authenticates (sets the cookies/tokens) into a login function that uses Playwright to visit a login page which is called from each test, Playwright offers something much better in that it can save browser storage state and re-use it. The idea is to login once and re-use the resulting browser state for every test that requires that role to work.

If the cookies/tokens don’t expire, you can capture them once, commit them to your code repository and simply re-use them:

First to capture:

  test('can view as admin', async ({ page, context }) => {
    await page.goto('http://webdriverjsdemo.github.io/auth/')
    await page.type('#firstname', 'Admin')
    await page.type('#surname', 'User')
    await page.click('#ok')
    await context.storageState({ path: 'storage/admin.json' });
    await page.goto('http://webdriverjsdemo.github.io/auth/')
    await expect(page.locator('text=Welcome name=AdminUser')).toBeVisible()
  })

  test('can view as standard user', async ({ page, context }) => {
    await page.goto('http://webdriverjsdemo.github.io/auth/')
    await page.type('#firstname', 'Standard')
    await page.type('#surname', 'Person')
    await page.click('#ok')
    await context.storageState({ path: 'storage/user.json' });
    await page.goto('http://webdriverjsdemo.github.io/auth/')
    await expect(page.locator('text=Welcome name=StandardPerson')).toBeVisible()
  })

And then to re-use the captured files:

import { test, expect } from '@playwright/test'

test.describe.parallel('Unauthenticated tests', () => {
  test('can view as guest', async ({ page }) => {
    await page.goto('http://webdriverjsdemo.github.io/auth/')
    await expect(page.locator('text=Hello Please Sign In')).toBeVisible()
  })
})

test.describe.parallel('Administrator tests', () => {
  test.use({storageState: './storage/admin.json'})
  test('can view as admin', async ({ page, context }) => {
    await page.goto('http://webdriverjsdemo.github.io/auth/')
    await expect(page.locator('text=Welcome name=AdminUser')).toBeVisible()
  })
})

test.describe.parallel('User tests', () => {
  test.use({storageState: './storage/user.json'})
  test('can view as standard user', async ({ page, context }) => {
    await page.goto('http://webdriverjsdemo.github.io/auth/')
    await expect(page.locator('text=Welcome name=StandardPerson')).toBeVisible()
  })
})

But what if like most apps your authentication cookies/tokens do expire?

Fortunately you can dynamically create the session state once per test run – in the global hooks – then simply refer to those same local storage files in each test.

You can create a file like global-setup.ts which generates our storage state files once per test run:

// global-setup.ts
import { Browser, chromium, FullConfig } from '@playwright/test'

async function globalSetup (config: FullConfig) {
  const browser = await chromium.launch()
  await saveStorage(browser, 'Standard', 'Person', 'storage/user.json')
  await saveStorage(browser, 'Admin', 'User', 'storage/admin.json')
  await browser.close()
}

async function saveStorage (browser: Browser, firstName: string, lastName: string, saveStoragePath: string) {
  const page = await browser.newPage()
  await page.goto('http://webdriverjsdemo.github.io/auth/')
  await page.type('#firstname', firstName)
  await page.type('#surname', lastName)
  await page.click('#ok')
  await page.context().storageState({ path: saveStoragePath })
}

export default globalSetup

which is referenced in playwright.config.ts

// playwright.config.ts
module.exports = {
  globalSetup: require.resolve('./global-setup'),
  reporter: [['list'], ['html']],
  retries: 0,
  use: {
    headless: true,
    screenshot: 'only-on-failure',
    video: 'retry-with-video',
    trace: 'on-first-retry'
  }
}

Once you have this set up our tests remain the same but the local storage values are captured and set once per test run:

test.describe.parallel('Admin tests', () => {
  test.use({ storageState: './storage/admin.json' })
  test('can view as admin', async ({ page, context }) => {
    await page.goto('http://webdriverjsdemo.github.io/auth/')
    await expect(page.locator('text=Welcome name=AdminUser')).toBeVisible()
  })
})

test.describe.parallel('User tests', () => {
  test.use({ storageState: './storage/user.json' })
  test('can view as standard user', async ({ page, context }) => {
    await page.goto('http://webdriverjsdemo.github.io/auth/')
    await expect(page.locator('text=Welcome name=StandardPerson')).toBeVisible()
  })
})

You can call test.use({ storageState: './storage/user.json' }) for a file or a test.describe block, so if all your tests in your test file use the same authentication role place it outside a test.describe block, otherwise place it within a test.describe block of tests that use the same authentication role. Different roles? Use different test.describe blocks with different test.use calls to different files in each.

What do you think of Playwright’s ability to capture and use browser storage state?

Categories
Playwright

Five reasons why Playwright is better than Cypress

It’s pretty obvious I am not a Cypress fan – heck I wrote a whole post about it 2 years ago.

I’ve since become very keen on Playwright and I thought I’d revisit Cypress to compare it with Playwright and see whether any of my long-standing gripes about Cypress have been addressed in 2+ years since I’ve used it.

Versions compared:

Cypress: 8.7.0 on Electron 93
Playwright: 1.16.0

Reason 1: Playwright is so much faster than Cypress

Up to 4x faster.

Take this simple example taken directly from the Cypress doco:

describe('My First Test', () => {
    it('clicking "type" shows the right headings', () => {
      cy.visit('https://example.cypress.io')
  
      cy.contains('type').click()
  
      // Should be on a new URL which includes '/commands/actions'
      cy.url().should('include', '/commands/actions')
  
      // Get an input, type into it and verify that the value has been updated
      cy.get('.action-email')
        .type('fake@email.com')
        .should('have.value', 'fake@email.com')
    })
})

Takes 8 seconds running on my M1 Macbook Air

 ✔  demo-spec.js                             00:08

I wrote the same example in Playwright:

import { test, expect } from '@playwright/test'

test.describe.parallel('My First Test', () => {
  test('clicking "type" shows the right headings', async ({ page }) => {
    await page.goto('https://example.cypress.io')
    await page.click('a:has-text("type")')
    // Should be on a new URL which includes '/commands/actions'
    await page.waitForURL(/.+\/commands\/actions$/)
    // Get an input, type into it and verify that the value has been updated
    await page.fill('.action-email', 'fake@email.com')
    await expect(page.locator('.action-email')).toHaveValue('fake@email.com')
  })
})

Running 1 test using 1 worker

  ✓  scenarios/compare.cypress.spec.ts:4:3 › My First Test › clicking "type" shows the right headings (1s)


  1 passed (2s)

4 times faster 😎

Reason 2: Playwright has first-class support for parallel running of tests on a single machine both locally and on CI without a subscription or account required

Cypress on the other hand requires you to set your tests to “Record” everything they do which sends everything to Cypress (or Sorry Cypress) if you want to run in parallel, and using a single machine is not recommended (despite it working perfectly in Playwright using my Macbook Air M1)

“While parallel tests can also technically run on a single machine, we do not recommend it since this machine would require significant resources to run your tests efficiently.”

https://docs.cypress.io/guides/guides/parallelization#Overview

Either this is true and Cypress is so bloated that running in parallel locally needs crazy expensive hardware, or the Cypress team want users to pay for their parallel execution service. Or both?

Reason 3: Playwright supports parallel tests within a single test file – to run Cypress tests in parallel you need to split them across files.

For example, in Playwright I can run these 8 tests in parallel:

import { test, expect } from '@playwright/test'

test.describe.parallel('My First Test', () => {
  test('clicking "type" shows the right headings 1', async ({ page }) => {
    await page.goto('https://example.cypress.io')
    await page.click('a:has-text("type")')
    await page.waitForURL(/.+\/commands\/actions$/)
    await page.fill('.action-email', 'fake@email.com')
    await expect(page.locator('.action-email')).toHaveValue('fake@email.com')
  })
  test('clicking "type" shows the right headings 2', async ({ page }) => {
    await page.goto('https://example.cypress.io')
    await page.click('a:has-text("type")')
    await page.waitForURL(/.+\/commands\/actions$/)
    await page.fill('.action-email', 'fake@email.com')
    await expect(page.locator('.action-email')).toHaveValue('fake@email.com')
  })
  test('clicking "type" shows the right headings 3', async ({ page }) => {
    await page.goto('https://example.cypress.io')
    await page.click('a:has-text("type")')
    await page.waitForURL(/.+\/commands\/actions$/)
    await page.fill('.action-email', 'fake@email.com')
    await expect(page.locator('.action-email')).toHaveValue('fake@email.com')
  })
  test('clicking "type" shows the right headings 4', async ({ page }) => {
    await page.goto('https://example.cypress.io')
    await page.click('a:has-text("type")')
    await page.waitForURL(/.+\/commands\/actions$/)
    await page.fill('.action-email', 'fake@email.com')
    await expect(page.locator('.action-email')).toHaveValue('fake@email.com')
  })
  test('clicking "type" shows the right headings 5', async ({ page }) => {
    await page.goto('https://example.cypress.io')
    await page.click('a:has-text("type")')
    await page.waitForURL(/.+\/commands\/actions$/)
    await page.fill('.action-email', 'fake@email.com')
    await expect(page.locator('.action-email')).toHaveValue('fake@email.com')
  })
  test('clicking "type" shows the right headings 6', async ({ page }) => {
    await page.goto('https://example.cypress.io')
    await page.click('a:has-text("type")')
    await page.waitForURL(/.+\/commands\/actions$/)
    await page.fill('.action-email', 'fake@email.com')
    await expect(page.locator('.action-email')).toHaveValue('fake@email.com')
  })
  test('clicking "type" shows the right headings 7', async ({ page }) => {
    await page.goto('https://example.cypress.io')
    await page.click('a:has-text("type")')
    await page.waitForURL(/.+\/commands\/actions$/)
    await page.fill('.action-email', 'fake@email.com')
    await expect(page.locator('.action-email')).toHaveValue('fake@email.com')
  })
  test('clicking "type" shows the right headings 8', async ({ page }) => {
    await page.goto('https://example.cypress.io')
    await page.click('a:has-text("type")')
    await page.waitForURL(/.+\/commands\/actions$/)
    await page.fill('.action-email', 'fake@email.com')
    await expect(page.locator('.action-email')).toHaveValue('fake@email.com')
  })
})

which takes only 6 seconds on my Macbook Air 😎

> npx playwright test "./scenarios/compare.cypress.bulk.spec.ts"

Running 8 tests using 4 workers

  ✓  scenarios/compare.cypress.bulk.spec.ts:4:3 › My First Test › clicking "type" shows the right headings 1 (4s)
  ✓  scenarios/compare.cypress.bulk.spec.ts:11:3 › My First Test › clicking "type" shows the right headings 2 (4s)
  ✓  scenarios/compare.cypress.bulk.spec.ts:18:3 › My First Test › clicking "type" shows the right headings 3 (4s)
  ✓  scenarios/compare.cypress.bulk.spec.ts:25:3 › My First Test › clicking "type" shows the right headings 4 (4s)
  ✓  scenarios/compare.cypress.bulk.spec.ts:46:3 › My First Test › clicking "type" shows the right headings 7 (1s)
  ✓  scenarios/compare.cypress.bulk.spec.ts:32:3 › My First Test › clicking "type" shows the right headings 5 (1s)
  ✓  scenarios/compare.cypress.bulk.spec.ts:39:3 › My First Test › clicking "type" shows the right headings 6 (1s)
  ✓  scenarios/compare.cypress.bulk.spec.ts:53:3 › My First Test › clicking "type" shows the right headings 8 (1s)

  8 passed (6s)

To possibly run these in parallel (using a server) in Cypress I’d need to structure these as individual files each with a single spec:

And to run them on the same machine I ran these 8 scenarios in Playwright in 6 seconds takes 45 seconds! 7.5 times slower!

Reason 4: Playwright fully supports async/await syntax for clean, readable code.

This is the Playwright code:

import { test, expect } from '@playwright/test'

test.describe.parallel('My First Test', () => {
  test('clicking "type" shows the right headings', async ({ page }) => {
    await page.goto('https://example.cypress.io')
    await page.click('a:has-text("type")')
    await page.waitForURL(/.+\/commands\/actions$/)
    await page.fill('.action-email', 'fake@email.com')
    await expect(page.locator('.action-email')).toHaveValue('fake@email.com')
  })
})

No chaining, no hard to understand code, no magic. You can see where test and expect come from, you can see where page comes from.

The same test in Cypress:

describe('My First Test', () => {
    it('clicking "type" shows the right headings', () => {
      cy.visit('https://example.cypress.io')
    
      cy.contains('type').click()
  
      // Should be on a new URL which includes '/commands/actions'
      cy.url().should('include', '/commands/actions')
  
      // Get an input, type into it and verify that the value has been updated
      cy.get('.action-email')
        .type('fake@email.com')
        .should('have.value', 'fake@email.com')
    })
})

Note how the function calls chain together – in this simple example it’s not too bad but it quickly gets out of hand. Also note there’s no indication that this is synchronous code – no await on any calls. Also where does this cy thing come from – oh yes that’s magic. Same for describe, again magic. No imports but lots of magic.

Reason 5: Playwright doesn’t need plugins

There’s so many limitations in Cypress there’s pretty much a plugin for everything. And there’s still so may trade-offs that can’t be be fixed by a plugin.

Need to send the tab key to the browser? There’s a plugin for that. And it doesn’t work well at all according to the comments on the Github issue for lack of tab key support open since 2016 (and still open since I last wrote about Cypress).

Need to upload a file? There’s also a plugin for that.

Need to run tests n-times in a row to measure repeatability? There’s a plugin for that.

Fortunately Playwright supports all of these things, and many more, natively. No plug-ins needed 😎

Summary

As you can see I’m still not a Cypress fan all these years later 😀

It’s just that now there’s a tool which is so much better and freely available to use.

So I’ll stand by my summary from 2 years ago which I can repeat here:

From a distance Cypress looks like a polished tool for automated testing – I just think it’s incorrectly marketed as an end-to-end testing tool when it’s really only good for component testing. There are too many limitations in the tool in acting like a real user to use it to create true end-to-end automated tests.

https://alisterbscott.com/2019/07/06/my-thoughts-on-cypress-io/

Categories
Playwright

API Testing using Playwright

I’m a big fan of being able to call APIs during e2e tests to set up data, verify things, do things and all that jazz.

Up until now I’ve been mixing in Supertest with Playwright to call APIs within tests, until now…

Playwright 1.16 includes the ability to call APIs both independently and using the page browser object (which sends the currently stored cookies for API requests).

I’ve updated my example TypeScript project to include these API calls directly using both page and request which looks like:

test('can POST a REST API and check response using approval style', async ({ request }) => {
    const response = await request.post('https://my-json-server.typicode.com/webdriverjsdemo/webdriverjsdemo.github.io/posts', { data: { title: 'Post 4' } })
    expect(response.status()).toBe(201)
    const body = await response.text()
    expect(body).toMatchSnapshot('post4.txt')
  })

  test('can POST a REST API and check response using assertion style (using page)', async ({ page }) => {
    const response = await page.request.post('https://my-json-server.typicode.com/webdriverjsdemo/webdriverjsdemo.github.io/posts', { data: { title: 'Post 4' } })
    expect(response.status()).toBe(201)
    const body = JSON.parse(await response.text())
    expect(body.id).toBe(4)
    expect(body.title).toBe('Post 4')
  })

It’s nice to be able to remove another dependency from my e2e tests and still allow calling APIs using a nice API and reusing the existing objects that Playwright provides. I am continually blown away by how Playwright continues getting more and more awesome 😎

Categories
Playwright

Selecting hard to identify elements with Playwright

Say you have a web app which displays a table of dynamic data that looks like this:

Admire my beautiful design skills!

And the DOM looks like:

<table>
    <theader>
      <tr>
        <th>Status</th>
        <th>Name</th>
        <th>Actions</th>
      </tr>
    </theader>
    <tbody>
      <tr>
        <td>OPEN</td>
        <td>Cactus</td>
        <td>
          <div>
            <button>CLICK ME</button>
          </div>
        </td>
      </tr>
      <tr>
        <td>CLOSED</td>
        <td>Succulent</td>
        <td>
          <div>
            <button>CLICK ME</button>
          </div>
        </td>
      </tr>
      <tr>
        <td>OPEN</td>
        <td>Aloe</td>
        <td>
          <div>
            <button>CLICK ME</button>
          </div>
        </td>
      </tr>
      <tr>
        <td>OPEN</td>
        <td>Agave</td>
        <td>
          <div>
            <button>CLICK ME</button>
          </div>
        </td>
      </tr>
    </tbody>
  </table>

Say you’re writing an e2e test that adds an item to the table dynamically (with a new name each time) and you need to click the button on the corresponding line to complete your e2e test.

How do you select the button based upon the name of the item you just created? Let’s save Agave?

I feel like the best possible approach is to add a data- attribute to the buttons so you can know exactly which one to click. Something like <button data-name="Agave">Agave</button"

A previous article I wrote on using data attributes

But if you can’t add that value to each row – what’s the next best thing to do?

Playwright has a pretty neat test recorder (even though I don’t like test recorders) – so let’s see what it generates

I launch Playwright Inspector (which has the Record function) with

npx playwright open https://webdriverjsdemo.github.io/table/

After clicking Record, and then clicking the four buttons on the table in order from top-bottom, the Inspector tool gives me:

Clicking the four buttons here
  await page.click('text=CLICK ME');
  await page.click('text=Succulent CLOSED CLICK ME >> button');
  await page.click('text=Aloe OPEN CLICK ME >> button');
  await page.click('text=Agave OPEN CLICK ME >> button');

The issue with this generated code is it relies on the status values which can change, also if the status column was removed in a future change it would fail which makes the tests brittle.

We can use the Playwright Inspector “Explore” feature to enter a selector value and see whether it finds what we’re looking for in the DOM.

If we took one of the selectors above and tried to remove the bit we don’t care about from the text you can see it doesn’t find what we want:

text=Agave >> button

Doing some research I found there’s an option to find any element containing some text, we can choose the row containing the text “Agave” (or whatever the name of the record we created is) and then choose the button belonging to that row:

tr:has-text("Agave") >> button

The important thing to note here is we are selecting the row, even though the cell contains the text we’re looking for.

I think this is probably the best option if we don’t have ability to change the DOM and add data attributes – it’s the least brittle as any additional columns could be added to the table and it would still work.

Some other options I investigated were locators based on position

For example, you can write a selector that finds the buttons to the right of the text we’re looking for:

button:right-of(:text("Agave"))

But as you can see from the image above it locates all the buttons as they’re all technically to the right of the text – just not immediately to the right.

Summary

Whilst it’s good practice to add testability features to your app, Playwright also offers powerful ways to select elements based on text and DOM layout if you can’t. You can read all about the Playwright selectors here.

Categories
Playwright

Running parallel Playwright Tests within a single spec file

Playwright by default runs all tests within a single spec file in order using the same worker, and runs tests in different spec files in parallel.

You can now use the test.describe.parallel block to specify that all tests contained in the block are independent and can be run in parallel. This can mean faster test runs.

There’s also test.describe.serial to specify the tests contained should be run one after another, and the subsequent tests won’t be run if an earlier test fails.

There’s also test.step which can be used to break tests down further into a series of steps.

I’ve updated my example TypeScript tests to use the parallel format.

Categories
Aside

Microsoft Teams

Most of my meetings nowadays are held online.

I hate being late to meetings, but I also like getting things done. When my calendar gives me the reminder about a meeting 10 mins before I’ve developed a habit of joining the meeting straight away so I am already in the meeting when other people join. That way I can continue my work without forgetting to join right on time. I’ve found if I don’t do that I’ll dismiss the notification, get caught up in my work (flow) and forget to join.

This was all well and good at my previous company where we used Google Meet. But my current company uses Microsoft Teams and when I join 10 mins early everyone else invited to the meeting gets a notification telling them I’ve started the meeting! Who designed Microsoft Teams thinking that was a good idea!?!

Any ideas how I can work with this?