如何在Protractor Cucumber中跳过失败测试并继续执行后续场景

Absolutely, there are straightforward ways to make your test suite keep running even when individual steps or scenarios fail—this is what’s commonly referred to as disabling "fail-fast" behavior, and almost every major testing framework supports this with simple tweaks or built-in features. Here’s how to do it across popular tools:

1. pytest (Python) #

• By default, pytest won’t stop on the first failure unless you use the -x/--exitfirst flag. Just omit that flag, and it’ll run all tests regardless of failures.
• If you want to keep running but keep failure output clean, use flags like --tb=short (concise tracebacks) or --no-header (reduces clutter).
• For individual test steps (where a single assertion failure stops the whole test function), use the pytest-check plugin. It lets you run multiple assertions and collect all failures at the end:

  # First install: pip install pytest-check
  import pytest_check as check
  
  def test_multi_step_workflow():
      check.equal(1, 1)  # Passes
      check.equal(2, 3)  # Fails, but test continues
      check.true(True)   # Runs after the failure
  

2. JUnit 5 (Java) #

• JUnit 5’s default behavior is to not stop the suite on individual test failures. If your setup is stopping early, check if you’ve set failFast=true on your @Test annotations—set it to false instead.
• To disable fail-fast at the suite level, add a configuration parameter to your suite class:

  import org.junit.platform.suite.api.ConfigurationParameter;
  import org.junit.platform.suite.api.Suite;
  import static org.junit.platform.engine.discovery.DiscoverySelectors.selectPackages;
  
  @Suite
  @SelectPackages("com.yourcompany.tests")
  @ConfigurationParameter(key = "junit.jupiter.execution.fail-fast.enabled", value = "false")
  public class TestSuite {}
  

• For multi-step tests where you want individual assertions to not halt the test, use assertAll():

  @Test
  void testMultipleAssertions() {
      assertAll(
          () -> assertEquals(1, 1),
          () -> assertEquals(2, 3), // Fails, but other assertions run
          () -> assertTrue(true)
      );
  }
  

3. TestNG (Java) #

• TestNG runs all tests by default even if some fail. If your suite is stopping early, check your suite XML for failOnTestFailure="true"—change it to false:

  <suite name="MyTestSuite" failOnTestFailure="false">
      <test name="CoreTests">
          <classes>
              <class name="com.yourcompany.tests.LoginTests"/>
              <class name="com.yourcompany.tests.CheckoutTests"/>
          </classes>
      </test>
  </suite>
  

• For individual test methods with multiple steps, use SoftAssert instead of Assert. It collects failures and reports them all at the end of the test:

  import org.testng.asserts.SoftAssert;
  
  @Test
  void testCheckoutFlow() {
      SoftAssert softAssert = new SoftAssert();
      softAssert.assertEquals(cart.getTotal(), 50.0);
      softAssert.assertEquals(checkout.getShippingCost(), 5.0); // Fails, test continues
      softAssert.assertTrue(checkout.isOrderConfirmed());
      softAssert.assertAll(); // Reports all collected failures
  }
  

4. Cypress (JavaScript) #

• Cypress stops on the first failure by default. To disable this at the suite level, add a listener to catch failures and prevent execution from halting:

  describe("E-Commerce Tests", () => {
      beforeEach(() => {
          // Prevent any failure from stopping the entire suite
          Cypress.on("fail", (err, runnable) => {
              console.log(`Test failed but continuing: ${err.message}`);
              return false;
          });
      });
  
      it("loads homepage", () => {
          cy.visit("/");
          cy.get("#non-existent-element").click(); // Fails, but next test runs
      });
  
      it("navigates to cart", () => {
          cy.get("#cart-button").click(); // Runs after the first test fails
      });
  });
  

The core takeaway is either disabling your framework’s built-in fail-fast behavior at the suite level, or using framework-specific tools (like soft assertions, plugins, or event handlers) to isolate individual failures so they don’t halt the entire test run.

内容的提问来源于stack exchange，提问作者Bhargav