This is a viewer only at the moment see the article on how this works.
To update the preview hit Ctrl-Alt-R (or ⌘-Alt-R on Mac) or Enter to refresh. The Save icon lets you save the markdown file to disk
This is a preview from the server running through my markdig pipeline
Thursday, 27 November 2025
Playwright es la solución oficial de Microsoft para la automatización de navegadores modernos, que ofrece soporte multi-navegador (Chrome, Firefox, Safari) con una sola API, depuración de trazas integrada, y emulación móvil. Esta guía cubre todo, desde la configuración hasta los patrones de prueba avanzados, comparándolo con PuppeteerSharp para ayudarle a elegir la herramienta correcta. TitiriteroSharp en su lugar - pero si necesita pruebas completas de cross-browser, Playwright es el camino a seguir.
Si has leído mi artículo sobre TitiriteroSharp, usted sabrá que soy un fan de las modernas herramientas de prueba E2E que no te hacen querer rasgar el pelo hacia fuera . TitiriteroSharp es brillante para las pruebas sólo Chrome , pero lo que si usted necesita para probar a través de múltiples navegadores ? Ahí es donde Playwright Pasa.
Playwright es la respuesta de Microsoft al problema de automatización del navegador, y han aprendido de todo lo que vino antes. Es como el hermano más ambicioso de PuppeteerSharp - hace todo lo que Tippeteer hace, pero añade Firefox y Safari soporte, mejor auto-esperar, y una gran cantidad de herramientas de depuración que hacen que encontrar problemas sea un problema absoluto.
En este artículo, te mostraré cómo utilizar Playwright para .NET para probar a través de Chrome, Firefox y Safari, con ejemplos de código real y patrones prácticos que puede utilizar hoy en día.
No me malinterpretes TitiriteroSharp es excelente si usted sólo necesita Chrome. Pero aquí es cuando Playwright tiene sentido:
Sus usuarios no todos utilizan Chrome. Utilizan:
Un error que sólo aparece en Safari puede perder el 20% de sus usuarios potenciales. Playwright le permite probar los tres con la misma API.
graph TB
subgraph "PuppeteerSharp"
PS[Chrome Only]
PS --> PS1[Fast setup]
PS --> PS2[Simple API]
PS --> PS3[Chrome DevTools]
end
subgraph "Playwright"
PW[Multi-Browser]
PW --> PW1[Chrome, Firefox, Safari]
PW --> PW2[Auto-waiting built-in]
PW --> PW3[Trace viewer]
PW --> PW4[Codegen tool]
PW --> PW5[Better error messages]
end
style PW stroke:#00aa00,stroke-width:3px
style PW1 stroke:#0066cc,stroke-width:2px
style PW3 stroke:#0066cc,stroke-width:2px
style PW4 stroke:#0066cc,stroke-width:2px
Instalar la Microsoft.Playwright Paquete NuGet:
dotnet add package Microsoft.Playwright
dotnet add package Microsoft.Playwright.NUnit # Or use xUnit
A continuación, instale los navegadores (estas descargas Cromo, Firefox, y WebKit):
pwsh bin/Debug/net9.0/playwright.ps1 install
O en Linux/Mac:
playwright install
Nota: El primer navegador instalar descargas alrededor de 400 MB. Las actualizaciones posteriores son mucho más pequeñas.
Aquí está mi configuración de proyecto de prueba:
<PackageReference Include="Microsoft.Playwright" Version="1.48.0" />
<PackageReference Include="Microsoft.Playwright.NUnit" Version="1.48.0" />
<PackageReference Include="xunit" Version="2.9.3" />
<PackageReference Include="xunit.runner.visualstudio" Version="3.1.4">
<PrivateAssets>all</PrivateAssets>
<IncludeAssets>runtime; build; native; contentfiles; analyzers; buildtransitive</IncludeAssets>
</PackageReference>
Estoy usando xUnit, pero el dramaturgo funciona igual de bien con NUnit o MSTest. De hecho, Playwright tiene un dedicado Integración de NUnit con ayudantes adicionales.
Similar a TitiriteroSharp, pero con soporte multi-navegador:
using Microsoft.Playwright;
using Xunit.Abstractions;
namespace Mostlylucid.Test.E2E;
public abstract class PlaywrightTestBase : IAsyncLifetime
{
protected IPlaywright Playwright = null!;
protected IBrowser Browser = null!;
protected IBrowserContext Context = null!;
protected IPage Page = null!;
protected readonly ITestOutputHelper Output;
protected const string BaseUrl = "http://localhost:8080";
// Override this in derived classes to test different browsers
protected virtual BrowserType BrowserType => BrowserType.Chromium;
protected PlaywrightTestBase(ITestOutputHelper output)
{
Output = output;
}
public async Task InitializeAsync()
{
// Create Playwright instance
Playwright = await Microsoft.Playwright.Playwright.CreateAsync();
// Launch the specified browser
Browser = await LaunchBrowserAsync();
// Create a new context (like an incognito window)
Context = await Browser.NewContextAsync(new BrowserNewContextOptions
{
ViewportSize = new ViewportSize { Width = 1400, Height = 900 },
IgnoreHTTPSErrors = true
});
// Create a new page
Page = await Context.NewPageAsync();
// Set default timeout
Page.SetDefaultTimeout(30000);
}
private async Task<IBrowser> LaunchBrowserAsync()
{
var options = new BrowserTypeLaunchOptions
{
Headless = true, // Set to false for debugging
};
return BrowserType switch
{
BrowserType.Chromium => await Playwright.Chromium.LaunchAsync(options),
BrowserType.Firefox => await Playwright.Firefox.LaunchAsync(options),
BrowserType.Webkit => await Playwright.Webkit.LaunchAsync(options),
_ => throw new ArgumentException($"Unknown browser type: {BrowserType}")
};
}
public async Task DisposeAsync()
{
await Page?.CloseAsync()!;
await Context?.CloseAsync()!;
await Browser?.CloseAsync()!;
Playwright?.Dispose();
}
}
public enum BrowserType
{
Chromium,
Firefox,
Webkit
}
Note que creamos un Context antes de crear una PageEste es un concepto de dramaturgo que TitiriteroSharp no tiene:
Los contextos le permiten:
// Navigation with better waiting
protected async Task NavigateAsync(string path)
{
var url = path.StartsWith("http") ? path : $"{BaseUrl}{path}";
await Page.GotoAsync(url, new PageGotoOptions
{
WaitUntil = WaitUntilState.NetworkIdle
});
}
// Playwright has better built-in waiting, but these are still useful
protected async Task<IElementHandle?> WaitForSelectorAsync(string selector)
{
try
{
return await Page.WaitForSelectorAsync(selector, new PageWaitForSelectorOptions
{
State = WaitForSelectorState.Visible
});
}
catch (TimeoutException)
{
return null;
}
}
// Improved element operations with auto-waiting
protected async Task<bool> ElementExistsAsync(string selector)
{
return await Page.Locator(selector).CountAsync() > 0;
}
protected async Task<string?> GetTextContentAsync(string selector)
{
var locator = Page.Locator(selector);
return await locator.TextContentAsync();
}
protected async Task TypeAsync(string selector, string text)
{
await Page.Locator(selector).FillAsync(text);
}
protected async Task ClickAsync(string selector)
{
await Page.Locator(selector).ClickAsync();
}
Nótese que Page.Locator() Esto es diferente de las llamadas de PuppeteerSharp. QuerySelector. Los localizadores son:
// PuppeteerSharp way (manual waiting)
await Page.WaitForSelectorAsync("#button");
await Page.ClickAsync("#button");
// Playwright way (auto-waiting)
await Page.Locator("#button").ClickAsync(); // Waits automatically!
Aquí es donde brilla Playwright. Puedes ejecutar la misma prueba en todos los navegadores:
public class CrossBrowserTests : PlaywrightTestBase
{
public CrossBrowserTests(ITestOutputHelper output) : base(output) { }
[Theory]
[InlineData(BrowserType.Chromium)]
[InlineData(BrowserType.Firefox)]
[InlineData(BrowserType.Webkit)]
public async Task HomePage_LoadsCorrectly_InAllBrowsers(BrowserType browserType)
{
// Override the browser type for this test
BrowserType = browserType;
await InitializeAsync();
// Arrange & Act
await NavigateAsync("/");
// Assert
var title = await Page.TitleAsync();
Assert.Contains("Mostlylucid", title);
Output.WriteLine($"✓ Test passed in {browserType}");
}
}
O cree clases de prueba separadas para cada navegador:
public class ChromiumTests : PlaywrightTestBase
{
protected override BrowserType BrowserType => BrowserType.Chromium;
public ChromiumTests(ITestOutputHelper output) : base(output) { }
[Fact]
public async Task FilterBar_WorksInChrome()
{
await NavigateAsync("/blog");
await Page.Locator("#LanguageDropDown button").ClickAsync();
await Page.Locator("#LanguageDropDown li:has-text('Spanish')").ClickAsync();
var url = Page.Url;
Assert.Contains("language=es", url);
}
}
public class FirefoxTests : PlaywrightTestBase
{
protected override BrowserType BrowserType => BrowserType.Firefox;
public FirefoxTests(ITestOutputHelper output) : base(output) { }
[Fact]
public async Task FilterBar_WorksInFirefox()
{
// Same test, different browser!
await NavigateAsync("/blog");
await Page.Locator("#LanguageDropDown button").ClickAsync();
await Page.Locator("#LanguageDropDown li:has-text('Spanish')").ClickAsync();
var url = Page.Url;
Assert.Contains("language=es", url);
}
}
public class SafariTests : PlaywrightTestBase
{
protected override BrowserType BrowserType => BrowserType.Webkit;
public SafariTests(ITestOutputHelper output) : base(output) { }
[Fact]
public async Task FilterBar_WorksInSafari()
{
await NavigateAsync("/blog");
await Page.Locator("#LanguageDropDown button").ClickAsync();
await Page.Locator("#LanguageDropDown li:has-text('Spanish')").ClickAsync();
var url = Page.Url;
Assert.Contains("language=es", url);
}
}
Veamos ejemplos reales que muestran las ventajas de Playwright:
[Fact]
public async Task SortingDropdown_UpdatesContentViaHTMX()
{
// Arrange
await NavigateAsync("/blog");
// Get first post title before sorting
var firstPostBefore = await Page.Locator("article h2 a").First.TextContentAsync();
Output.WriteLine($"First post before: {firstPostBefore}");
// Act - Change sort order
await Page.Locator("#orderSelect").SelectOptionAsync("date_asc");
// Playwright automatically waits for HTMX to update the DOM!
await Page.WaitForLoadStateAsync(LoadState.NetworkIdle);
// Assert - Content should have changed
var firstPostAfter = await Page.Locator("article h2 a").First.TextContentAsync();
Output.WriteLine($"First post after: {firstPostAfter}");
var selectValue = await Page.Locator("#orderSelect").InputValueAsync();
Assert.Equal("date_asc", selectValue);
}
Manijas de reproducción Alpine.js reactividad y animaciones sin problemas:
[Fact]
public async Task AlpineDropdown_OpensAndCloses()
{
await NavigateAsync("/blog");
// Click to open dropdown (Alpine.js controlled)
await Page.Locator("#LanguageDropDown button").ClickAsync();
// Wait for Alpine.js animation
await Page.WaitForSelectorAsync("#LanguageDropDown div[x-show]", new()
{
State = WaitForSelectorState.Visible
});
// Check dropdown is visible
var isVisible = await Page.Locator("#LanguageDropDown div[x-show]").IsVisibleAsync();
Assert.True(isVisible);
// Click outside to close
await Page.Locator("body").ClickAsync(new LocatorClickOptions
{
Position = new Position { X = 0, Y = 0 }
});
// Verify closed
await Page.WaitForSelectorAsync("#LanguageDropDown div[x-show]", new()
{
State = WaitForSelectorState.Hidden
});
isVisible = await Page.Locator("#LanguageDropDown div[x-show]").IsVisibleAsync();
Assert.False(isVisible);
}
[Fact]
public async Task CommentForm_ValidatesRequiredFields()
{
await NavigateAsync("/blog/some-post");
// Try to submit without filling fields
await Page.Locator("#comment-submit").ClickAsync();
// Check HTML5 validation messages
var nameInput = Page.Locator("#comment-name");
var validationMessage = await nameInput.EvaluateAsync<string>("el => el.validationMessage");
Assert.NotEmpty(validationMessage);
Output.WriteLine($"Validation message: {validationMessage}");
// Fill form properly
await nameInput.FillAsync("Test User");
await Page.Locator("#comment-email").FillAsync("[email protected]");
await Page.Locator("#comment-content").FillAsync("Great article!");
// Submit
await Page.Locator("#comment-submit").ClickAsync();
// Wait for success
await Page.Locator(".comment-success").WaitForAsync();
}
[Fact]
public async Task PageLoadsWithoutImages_ToTestPerformance()
{
// Block all image requests
await Page.RouteAsync("**/*.{png,jpg,jpeg,gif,webp}", route => route.AbortAsync());
await NavigateAsync("/blog");
// Page should still function without images
var title = await Page.Locator("h1").TextContentAsync();
Assert.NotEmpty(title);
}
[Fact]
public async Task MockAPIResponse_ForTesting()
{
// Intercept API calls and return mock data
await Page.RouteAsync("**/api/posts", route => route.FulfillAsync(new()
{
Status = 200,
ContentType = "application/json",
Body = System.Text.Json.JsonSerializer.Serialize(new
{
posts = new[]
{
new { id = 1, title = "Mock Post 1" },
new { id = 2, title = "Mock Post 2" }
}
})
}));
await NavigateAsync("/blog");
// Should show mock data
var posts = await Page.Locator(".post-title").CountAsync();
Assert.Equal(2, posts);
}
public async Task InitializeAsync()
{
Playwright = await Microsoft.Playwright.Playwright.CreateAsync();
Browser = await LaunchBrowserAsync();
// Start tracing
Context = await Browser.NewContextAsync();
await Context.Tracing.StartAsync(new()
{
Screenshots = true,
Snapshots = true,
Sources = true
});
Page = await Context.NewPageAsync();
}
public async Task DisposeAsync()
{
// Save trace on test completion
var tracePath = Path.Combine("traces", $"{TestContext.TestName}.zip");
await Context.Tracing.StopAsync(new()
{
Path = tracePath
});
// View trace with: playwright show-trace traces/TestName.zip
await Page?.CloseAsync()!;
await Context?.CloseAsync()!;
await Browser?.CloseAsync()!;
Playwright?.Dispose();
}
A continuación, ver el rastro:
playwright show-trace traces/TestName.zip
Esto abre una interfaz de usuario mostrando:
Es absolutamente brillante para la depuración.
public async Task DisposeAsync()
{
// Take screenshot if test failed
if (TestContext.CurrentTestOutcome != TestOutcome.Passed)
{
var screenshot = await Page.ScreenshotAsync();
File.WriteAllBytes($"failure-{TestContext.TestName}.png", screenshot);
Output.WriteLine($"Screenshot saved: failure-{TestContext.TestName}.png");
}
await Page?.CloseAsync()!;
await Context?.CloseAsync()!;
await Browser?.CloseAsync()!;
Playwright?.Dispose();
}
[Fact]
public async Task BlogPage_WorksOnMobile()
{
// Create context with mobile emulation
var iPhone = Playwright.Devices["iPhone 13"];
await using var context = await Browser.NewContextAsync(iPhone);
await using var page = await context.NewPageAsync();
await page.GotoAsync($"{BaseUrl}/blog");
// Check mobile menu is visible
var mobileMenu = page.Locator(".mobile-menu");
await Expect(mobileMenu).ToBeVisibleAsync();
// Desktop menu should be hidden
var desktopMenu = page.Locator(".desktop-menu");
await Expect(desktopMenu).Not.ToBeVisibleAsync();
}
El dramaturgo viene con descriptores de dispositivos donde dice:
[Fact]
public async Task DarkMode_TogglesCorrectly()
{
// Start with dark color scheme
await using var context = await Browser.NewContextAsync(new()
{
ColorScheme = ColorScheme.Dark
});
await using var page = await context.NewPageAsync();
await page.GotoAsync($"{BaseUrl}");
// Check dark mode is active
var isDark = await page.EvaluateAsync<bool>(
"() => window.matchMedia('(prefers-color-scheme: dark)').matches"
);
Assert.True(isDark);
// Check background color reflects dark mode
var bgColor = await page.Locator("body").EvaluateAsync<string>(
"el => getComputedStyle(el).backgroundColor"
);
Assert.Contains("rgb(0, 0, 0)", bgColor); // Dark background
}
Aquí está la misma prueba en ambas bibliotecas para mostrar las diferencias:
[Fact]
public async Task FilterTest_PuppeteerSharp()
{
await Page.GoToAsync("http://localhost:8080/blog");
// Manual waiting required
await Page.WaitForSelectorAsync("#LanguageDropDown button");
await Page.ClickAsync("#LanguageDropDown button");
// Wait for dropdown animation
await Task.Delay(300);
// Click Spanish option
await Page.WaitForSelectorAsync("#LanguageDropDown li:nth-child(2) a");
await Page.ClickAsync("#LanguageDropDown li:nth-child(2) a");
// Wait for navigation
await Task.Delay(500);
// Check URL
var url = Page.Url;
Assert.Contains("language=", url);
}
[Fact]
public async Task FilterTest_Playwright()
{
await Page.GotoAsync("http://localhost:8080/blog");
// Auto-waiting built in
await Page.Locator("#LanguageDropDown button").ClickAsync();
// Click Spanish option (waits automatically for visibility)
await Page.Locator("#LanguageDropDown li:has-text('Spanish')").ClickAsync();
// Check URL (waits automatically for navigation)
await Expect(Page).ToHaveURLAsync(new Regex(".*language=.*"));
}
Aviso:
WaitForSelectorAsync NecesarioTask.Delay NecesarioExpecthas-text selectorComo PuppeteerSharp, Playwright puede generar PDFs. La API es casi idéntica:
public async Task<byte[]> GeneratePdfAsync(string url)
{
await using var browser = await Playwright.Chromium.LaunchAsync();
await using var page = await browser.NewPageAsync();
await page.GotoAsync(url);
return await page.PdfAsync(new()
{
Format = "A4",
PrintBackground = true,
Margin = new()
{
Top = "20mm",
Right = "20mm",
Bottom = "20mm",
Left = "20mm"
}
});
}
Lo mismo PDF gotchas del artículo de PuppeteerSharp aplicar aquí también.
graph TD
A[Need E2E Testing?] --> B{Multi-browser required?}
B -->|Yes| C[Playwright]
B -->|No| D{Chrome only?}
D -->|Yes| E[PuppeteerSharp or Playwright]
D -->|No| C
C --> F[Benefits]
F --> F1[Test all major browsers]
F --> F2[Better debugging tools]
F --> F3[More reliable waiting]
F --> F4[Mobile emulation]
E --> G[PuppeteerSharp Benefits]
G --> G1[Simpler setup]
G --> G2[Smaller overhead]
G --> G3[Chromium focused]
style C stroke:#00aa00,stroke-width:3px
style E stroke:#0066cc,stroke-width:2px
Elija Playwright cuando:
Elija PuppeteerSharp cuando:
Playwright funciona perfectamente en tuberías CI/CD. Aquí está un completo Acciones de GitHub flujo de trabajo:
name: Playwright Tests
on:
push:
branches: [ main ]
pull_request:
branches: [ main ]
jobs:
test:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v3
- name: Setup .NET
uses: actions/setup-dotnet@v3
with:
dotnet-version: '9.0.x'
- name: Install dependencies
run: dotnet restore
- name: Build
run: dotnet build --no-restore
- name: Install Playwright browsers
run: pwsh Mostlylucid.Test/bin/Debug/net9.0/playwright.ps1 install --with-deps
- name: Start application
run: |
dotnet run --project Mostlylucid/Mostlylucid.csproj &
echo $! > app.pid
- name: Wait for application
run: |
timeout 60 bash -c 'until curl -f http://localhost:8080/health; do sleep 2; done'
- name: Run Playwright tests
run: |
dotnet test Mostlylucid.Test/Mostlylucid.Test.csproj \
--filter "Category=E2E" \
--logger "console;verbosity=detailed"
- name: Upload test results
if: always()
uses: actions/upload-artifact@v3
with:
name: playwright-results
path: |
**/test-results/
**/traces/
- name: Stop application
if: always()
run: kill $(cat app.pid) || true
Ejecución de pruebas de reproducción en Docker requiere dependencias del sistema de instalación:
FROM mcr.microsoft.com/dotnet/sdk:9.0 AS build
# Install Playwright dependencies
RUN apt-get update && apt-get install -y \
libnss3 \
libnspr4 \
libatk1.0-0 \
libatk-bridge2.0-0 \
libcups2 \
libdrm2 \
libxkbcommon0 \
libxcomposite1 \
libxdamage1 \
libxfixes3 \
libxrandr2 \
libgbm1 \
libasound2
WORKDIR /app
COPY . .
RUN dotnet restore
RUN dotnet build
RUN pwsh Mostlylucid.Test/bin/Debug/net9.0/playwright.ps1 install
CMD ["dotnet", "test"]
De mis pruebas en este blog:
Característica TitiriteroSharp dramaturgo |---------|---------------|------------| | Velocidad de ensayo cromada ~2.5s ~2.8s ~2.8s ~2.8s ~2.5s ~2.5s ~2.5s ~2.8s ~2.8s ~2.8s ~2.8s ~2.8s ~2.8s ~2.8s ~2.5s ~2.5s ~2.5s ~2.8s ~2.8s ~2.8s ~2.8s ~2.8s ~2.8s ~2.8s | Multi-navegador ✓ (adiciones ~1s por navegador) | Fiabilidad de la espera automática Bueno (manual) Excelente (automático) | Memoria por navegador ~150MB ~180MB | Complejidad de configuración # Simple # # Moderado # | Herramientas de depuración DevTools Chrome Visor de rastros + DevTools | Emulación móvil # Básico # # Excelente #
Playwright es un poco más lento y utiliza un poco más de memoria, pero los beneficios de fiabilidad y depuración generalmente superan esto.
Los localizadores de Playwright son estrictos por defecto:
// This errors if multiple buttons exist
await Page.Locator("button").ClickAsync();
// Be specific
await Page.Locator("button#submit").ClickAsync();
// Or use .First if you really want the first match
await Page.Locator("button").First.ClickAsync();
No te olvides de correr. playwright install después de añadir el paquete. Los navegadores no están incluidos en el paquete NuGet.
Recuerde las pruebas de aislamiento de contextos:
// Bad - state leaks between tests
await Page.GotoAsync("/login");
// Do stuff...
await Page.GotoAsync("/dashboard");
// Previous login state might still exist
// Good - fresh context per test
await using var context = await Browser.NewContextAsync();
await using var page = await context.NewPageAsync();
// Completely isolated
Safari (WebKit) puede comportarse de manera diferente:
// May work in Chrome but fail in WebKit
await Page.Locator(".dropdown").Hover();
await Page.Locator(".dropdown-item").ClickAsync();
// More reliable across browsers
await Page.Locator(".dropdown").ClickAsync();
await Page.WaitForSelectorAsync(".dropdown-item");
await Page.Locator(".dropdown-item").ClickAsync();
Playwright representa la evolución de la automatización del navegador para los desarrolladores .NET. Si usted necesita soporte multi-navegador, es la opción clara. Sí, es ligeramente más complejo que TitiriteroSharp, pero los beneficios son sustanciales:
Mi recomendación:
El observador de trazas solo me ha ahorrado horas de depuración. Ser capaz de reproducir una prueba fallida y ver exactamente lo que el navegador vio en cada paso es invaluable.
Dale a Playwright una oportunidad en tu próximo proyecto - te sorprenderá gratamente lo fácil que es hacer pruebas de cross-browser.
© 2026 Scott Galloway — Unlicense — All content and source code on this site is free to use, copy, modify, and sell.