Enhanced JUnit XML Generator for XCTests and Swift Testing

tools/test_xml_generator/BUILD

@@ -0,0 +1,25 @@
+load("@rules_python//python:py_binary.bzl", "py_binary")
+load("@rules_python//python:py_library.bzl", "py_library")
+load("@rules_python//python:py_test.bzl", "py_test")
+load("//tools:binary_env.bzl", "binary_env")
+
+licenses(["notice"])
+
+py_binary(
+    name = "generate_test_xml",
+    srcs = ["generate_test_xml.py"],
+    env = binary_env.common_python_utf8_env(),
+    python_version = "PY3",
+    srcs_version = "PY3",
+    visibility = ["//visibility:public"],
+)
+
+py_library(
+    name = "generate_test_xml_lib",
+    srcs = [
+        "generate_test_xml.py",
+    ],
+    srcs_version = "PY3",
+    visibility = ["//visibility:public"],
+)
+

tools/test_xml_generator/README.md

@@ -0,0 +1,191 @@
+# Test XML Generator
+
+This directory contains tools for generating proper JUnit XML output from iOS and macOS test runs.
+
+## Overview
+
+When running iOS/macOS unit tests with Bazel, the default test output may not always produce proper JUnit XML format that CI systems expect. This tool provides a post-action script that:
+
+1. Parses the test log output from xcodebuild
+2. Extracts test results (passes, failures, timing)
+3. Generates proper JUnit XML format
+4. Writes the XML to the location Bazel expects (`$XML_OUTPUT_FILE`)
+
+## Components
+
+- **`generate_test_xml.py`**: Python script that parses test logs and generates JUnit XML
+- **`test_runners.bzl`**: Pre-configured test runners with XML generation enabled
+- **`BUILD`**: Bazel build definitions
+
+## Usage
+
+### Option 1: Use Pre-configured Test Runners
+
+The easiest way to use this tool is to use one of the pre-configured test runners:
+
+```python
+load("@build_bazel_rules_apple//apple:ios.bzl", "ios_unit_test")
+
+ios_unit_test(
+    name = "MyAppTests",
+    minimum_os_version = "15.0",
+    deps = [":MyAppTestsLib"],
+    runner = "//tools/test_xml_generator:ios_xctestrun_runner_with_xml",
+)
+```
+
+Available runners:
+- `//tools/test_xml_generator:ios_xctestrun_runner_with_xml` - For iOS tests using xcodebuild
+- `//tools/test_xml_generator:ios_test_runner_with_xml` - For iOS tests using custom runner
+- `//tools/test_xml_generator:macos_test_runner_with_xml` - For macOS tests
+
+### Option 2: Create Your Own Custom Runner
+
+You can create your own test runner with custom configuration:
+
+```python
+# In your BUILD file
+load(
+    "@build_bazel_rules_apple//apple/testing/default_runner:ios_xctestrun_runner.bzl",
+    "ios_xctestrun_runner",
+)
+
+ios_xctestrun_runner(
+    name = "my_custom_ios_runner",
+    post_action = "//tools/test_xml_generator:generate_test_xml",
+    post_action_determines_exit_code = False,  # Don't fail build if XML generation fails
+    # Add other custom configuration here
+    device_type = "iPhone 14",
+    os_version = "16.0",
+)
+
+# Use in your test
+load("@build_bazel_rules_apple//apple:ios.bzl", "ios_unit_test")
+
+ios_unit_test(
+    name = "MyAppTests",
+    minimum_os_version = "15.0",
+    deps = [":MyAppTestsLib"],
+    runner = ":my_custom_ios_runner",
+)
+```
+
+### Option 3: Add Post-Action to Existing Runner
+
+If you already have a custom runner, you can add the post-action:
+
+```python
+ios_xctestrun_runner(
+    name = "my_existing_runner",
+    post_action = "//tools/test_xml_generator:generate_test_xml",
+    post_action_determines_exit_code = False,
+    # ... your existing configuration ...
+)
+```
+
+## How It Works
+
+1. **Test Execution**: The test runner executes your tests normally
+2. **Log Capture**: Test output is captured to `$TEST_LOG_FILE` (typically in `/tmp`)
+3. **Post-Action**: After tests complete, the `generate_test_xml.py` script runs with these environment variables:
+   - `TEST_LOG_FILE`: Path to the test log
+   - `XML_OUTPUT_FILE`: Path where Bazel expects the XML output
+   - `TEST_EXIT_CODE`: Exit code from the test run
+   - `TEST_XCRESULT_BUNDLE_PATH`: Path to XCResult bundle (if available)
+   - `SIMULATOR_UDID`: Simulator ID (for iOS tests)
+4. **XML Generation**: The script parses the log and generates JUnit XML
+5. **Output**: XML is written to the correct location for Bazel and CI systems
+
+## Supported Test Formats
+
+The parser supports multiple test output formats:
+
+- **XCTest (Objective-C/Swift)**: Standard XCTest output format
+  ```
+  Test Case '-[MyTests testExample]' started.
+  Test Case '-[MyTests testExample]' passed (0.001 seconds).
+  ```
+
+- **Swift Testing (Xcode 16+)**: New Swift Testing framework
+  ```
+  Test MyTests/testExample started.
+  Test MyTests/testExample passed after 0.001 seconds.
+  ```
+
+## Output Format
+
+The generated XML follows the JUnit XML format:
+
+```xml
+<?xml version="1.0" ?>
+<testsuites name="iOS/macOS Tests" tests="10" failures="1" errors="0" time="2.5" timestamp="2024-01-01T12:00:00Z">
+  <testsuite name="MyAppTests" tests="10" failures="1" errors="0" time="2.5">
+    <testcase classname="MyAppTests" name="testExample" time="0.001"/>
+    <testcase classname="MyAppTests" name="testFailure" time="0.002">
+      <failure message="XCTAssertEqual failed" type="XCTestFailure">
+        MyAppTests.swift:42: XCTAssertEqual failed: ("expected") is not equal to ("actual")
+      </failure>
+      <system-err>
+        Test output and error messages...
+      </system-err>
+    </testcase>
+  </testsuite>
+</testsuites>
+```
+
+## Troubleshooting
+
+### No test cases found in log output
+
+If you see this warning, it means the parser couldn't find test results in the log. This could be because:
+
+1. Tests didn't run at all (build failure, configuration issue)
+2. Test output format is different than expected
+3. The log file is empty or corrupted
+
+Check the test log manually to see what format the output is in.
+
+### XML file not created
+
+Make sure:
+1. The runner has the post-action configured correctly
+2. The `//tools/test_xml_generator:generate_test_xml` target is accessible
+3. Check the test output for any error messages from the post-action script
+
+### Tests fail but XML shows all passed
+
+The post-action runs after tests complete but doesn't change the test exit code (unless `post_action_determines_exit_code = True`). The XML should reflect the actual test results from the log.
+
+## Development
+
+To test the XML generator locally:
+
+```bash
+# Run tests and generate XML
+bazel test //path/to:test --test_output=all
+
+# Check the generated XML
+cat bazel-testlogs/path/to/test/test.xml
+
+# Debug with verbose output
+TEST_LOG_FILE=/path/to/test.log \
+XML_OUTPUT_FILE=/tmp/test.xml \
+TEST_EXIT_CODE=0 \
+python3 tools/test_xml_generator/generate_test_xml.py
+```
+
+## Configuration Options
+
+### post_action_determines_exit_code
+
+By default, this is set to `False`, meaning XML generation failures won't fail your test. This is recommended because you still want to know if your tests failed even if XML generation has issues.
+
+Set to `True` if you want XML generation failures to fail the build:
+
+```python
+ios_xctestrun_runner(
+    name = "strict_runner",
+    post_action = "//tools/test_xml_generator:generate_test_xml",
+    post_action_determines_exit_code = True,  # Fail build if XML generation fails
+)
+```