Add --scan-dryrun-results for Maven subset to honor pom.xml test filtering

smart_tests/test_runners/maven.py

@@ -1,6 +1,7 @@
 import glob
 import os
 import re
+import xml.etree.ElementTree as ET
 from typing import Annotated, Dict, List
 
 import click
@@ -50,6 +51,29 @@ def is_file(f: str) -> bool:
     return False
 
 
+def parse_surefire_reports() -> List[Dict[str, str]]:
+    """Collect the test classes Surefire discovered from its TEST-*.xml reports.
+
+    Intended to be run against reports produced by a JUnit 5 dry run
+    (`mvn test -Djunit.platform.execution.dryRun.enabled=true`): Surefire applies
+    all of its own filtering (excludedGroups, includedGroups, excludes, profiles,
+    ...) before the dry run, so only the tests Maven would actually execute get a
+    report. Each report's root `name` attribute is the fully-qualified class name.
+    """
+    test_paths = []
+    report_pattern = os.path.join('**', 'target', 'surefire-reports', 'TEST-*.xml')
+    report_files = glob.glob(report_pattern, recursive=True)
+
+    for report_file in report_files:
+        try:
+            if classname := ET.parse(report_file).getroot().get('name'):
+                test_paths.append({"type": "class", "name": classname})
+        except ET.ParseError as e:
+            click.secho(f"Warning: Could not parse {report_file}: {e}", fg='yellow', err=True)
+
+    return test_paths
+
+
 @smart_tests.subset
 def subset(
     client: Subset,
@@ -68,6 +92,12 @@ def subset(
         help="Scan testCompile/default-testCompile/createdFiles.lst for *.lst files generated by `mvn compile` and "
              "use them as test inputs."
     )] = False,
+    is_scan_dryrun_results: Annotated[bool, typer.Option(
+        "--scan-dryrun-results",
+        help="Scan surefire reports generated by `mvn test -Djunit.platform.execution.dryRun.enabled=true`. "
+             "Only the tests Maven would actually run (after applying pom.xml excludedGroups/excludes/etc.) are "
+             "sent to the subset. Run the dry run before this command. JUnit 5 only."
+    )] = False,
     exclude_rules: Annotated[List[str], typer.Option(
         "--exclude",
         multiple=True,
@@ -108,6 +138,18 @@ def file2test(f: str) -> List | None:
     if source_roots is None:
         source_roots = []
 
+    if is_scan_dryrun_results:
+        tests = parse_surefire_reports()
+        if not tests:
+            raise BadCmdLineException(
+                "No surefire reports found under **/target/surefire-reports/. "
+                "Run `mvn test -Djunit.platform.execution.dryRun.enabled=true` before this command.")
+        for test in tests:
+            client.test_paths.append([test])
+        client.same_bin_formatter = lambda s: [{"type": "class", "name": s}]
+        client.run()
+        return
+
     files_to_read = list(test_compile_created_file)
     if is_scan_test_compile_lst:
         if len(test_compile_created_file) > 0:

tests/data/maven/dryrun-test/target/surefire-reports/TEST-com.example.demo.CalculatorTest.xml

@@ -0,0 +1,4 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<testsuite xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:noNamespaceSchemaLocation="https://maven.apache.org/surefire/maven-surefire-plugin/xsd/surefire-test-report.xsd" version="3.0.2" name="com.example.demo.CalculatorTest" time="0.004" tests="1" errors="0" skipped="1" failures="0">
+  <testcase name="add" classname="com.example.demo.CalculatorTest" time="0"/>
+</testsuite>

tests/data/maven/dryrun-test/target/surefire-reports/TEST-com.example.demo.UserTest.xml

@@ -0,0 +1,4 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<testsuite xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:noNamespaceSchemaLocation="https://maven.apache.org/surefire/maven-surefire-plugin/xsd/surefire-test-report.xsd" version="3.0.2" name="com.example.demo.UserTest" time="0.002" tests="1" errors="0" skipped="1" failures="0">
+  <testcase name="create" classname="com.example.demo.UserTest" time="0"/>
+</testsuite>

tests/data/maven/subset_scan_dryrun_results_result.json

@@ -0,0 +1,12 @@
+{
+  "testPaths": [
+    [{"type": "class", "name": "com.example.demo.CalculatorTest"}],
+    [{"type": "class", "name": "com.example.demo.UserTest"}]
+  ],
+  "testRunner": "maven",
+  "goal": {"type": "subset-by-percentage", "percentage": 0.1},
+  "ignoreNewTests": false,
+  "session": {"id": "16"},
+  "getTestsFromGuess": false,
+  "getTestsFromPreviousSessions": false
+}

tests/test_runners/test_maven.py

@@ -167,6 +167,36 @@ def test_subset_with_exclude(self):
         self.assert_success(result)
         self.assert_subset_payload('subset_with_exclude_rules_result.json')
 
+    @responses.activate
+    @mock.patch.dict(os.environ, {"SMART_TESTS_TOKEN": CliTestCase.smart_tests_token})
+    def test_scan_dryrun_results(self):
+        # Reports live under <dryrun-test>/target/surefire-reports/. The flag globs
+        # **/target/surefire-reports/TEST-*.xml relative to the cwd, so run from there.
+        test_data_dir = str(self.test_files_dir.joinpath('dryrun-test').resolve())
+        original_dir = os.getcwd()
+        try:
+            os.chdir(test_data_dir)
+            result = self.cli('subset', '--target', '10%', '--session',
+                              self.session, 'maven', '--scan-dryrun-results')
+            self.assert_success(result)
+            self.assert_subset_payload('subset_scan_dryrun_results_result.json')
+        finally:
+            os.chdir(original_dir)
+
+    @responses.activate
+    @mock.patch.dict(os.environ, {"SMART_TESTS_TOKEN": CliTestCase.smart_tests_token})
+    def test_scan_dryrun_results_no_reports(self):
+        original_dir = os.getcwd()
+        with tempfile.TemporaryDirectory() as temp_dir:
+            try:
+                os.chdir(temp_dir)
+                result = self.cli('subset', '--target', '10%', '--session',
+                                  self.session, 'maven', '--scan-dryrun-results')
+                self.assertNotEqual(result.exit_code, 0)
+                self.assertIn("No surefire reports found", result.output)
+            finally:
+                os.chdir(original_dir)
+
     def test_glob(self):
         for x in [
             'foo/BarTest.java',