[9.0] WIP (#121463)

test/framework/src/main/java/org/elasticsearch/test/AbstractXContentTestCase.java

@@ -145,8 +145,21 @@ private XContentTester(
         public void test() throws IOException {
             for (int runs = 0; runs < numberOfTestRuns; runs++) {
                 XContentType xContentType = randomFrom(XContentType.values()).canonical();
-                T testInstance = instanceSupplier.apply(xContentType);
+                T testInstance = null;
                 try {
+                    if (xContentType.equals(XContentType.YAML)) {
+                        testInstance = randomValueOtherThanMany(instance -> {
+                            // unicode character U+0085 (NEXT LINE (NEL)) doesn't survive YAML round trip tests (see #97716)
+                            // get a new random instance if we detect this character in the xContent output
+                            try {
+                                return toXContent.apply(instance, xContentType).utf8ToString().contains("\u0085");
+                            } catch (IOException e) {
+                                throw new RuntimeException(e);
+                            }
+                        }, () -> instanceSupplier.apply(xContentType));
+                    } else {
+                        testInstance = instanceSupplier.apply(xContentType);
+                    }
                     BytesReference originalXContent = toXContent.apply(testInstance, xContentType);
                     BytesReference shuffledContent = insertRandomFieldsAndShuffle(
                         originalXContent,
@@ -173,7 +186,9 @@ public void test() throws IOException {
                         dispose.accept(parsed);
                     }
                 } finally {
-                    dispose.accept(testInstance);
+                    if (testInstance != null) {
+                        dispose.accept(testInstance);
+                    }
                 }
             }
         }

test/framework/src/test/java/org/elasticsearch/test/AbstractXContentTestCaseTests.java

@@ -12,11 +12,13 @@
 import com.carrotsearch.randomizedtesting.RandomizedContext;
 
 import org.elasticsearch.common.bytes.BytesReference;
+import org.elasticsearch.xcontent.ToXContentFragment;
 import org.elasticsearch.xcontent.XContentBuilder;
 import org.elasticsearch.xcontent.XContentFactory;
 import org.elasticsearch.xcontent.XContentParser;
 import org.elasticsearch.xcontent.XContentType;
 
+import java.io.IOException;
 import java.util.Map;
 
 import static org.hamcrest.Matchers.equalTo;
@@ -49,4 +51,42 @@ public void testInsertRandomFieldsAndShuffle() throws Exception {
             assertThat(mapOrdered.keySet().iterator().next(), not(equalTo("field")));
         }
     }
+
+    private record TestToXContent(String field, String value) implements ToXContentFragment {
+
+        @Override
+        public XContentBuilder toXContent(XContentBuilder builder, Params params) throws IOException {
+            return builder.field(field, value);
+        }
+    }
+
+    public void testYamlXContentRoundtripSanitization() throws Exception {
+        var test = new AbstractXContentTestCase<TestToXContent>() {
+
+            @Override
+            protected TestToXContent createTestInstance() {
+                // we need to randomly create both a "problematic" and an okay version in order to ensure that the sanitization code
+                // can draw at least one okay version if polled often enough
+                return randomBoolean() ? new TestToXContent("a\u0085b", "def") : new TestToXContent("a b", "def");
+            }
+
+            @Override
+            protected TestToXContent doParseInstance(XContentParser parser) throws IOException {
+                assertEquals(XContentParser.Token.START_OBJECT, parser.nextToken());
+                assertEquals(XContentParser.Token.FIELD_NAME, parser.nextToken());
+                String name = parser.currentName();
+                assertEquals(XContentParser.Token.VALUE_STRING, parser.nextToken());
+                String value = parser.text();
+                assertEquals(XContentParser.Token.END_OBJECT, parser.nextToken());
+                return new TestToXContent(name, value);
+            };
+
+            @Override
+            protected boolean supportsUnknownFields() {
+                return false;
+            }
+        };
+        // testFromXContent runs 20 repetitions, enough to hit a YAML xcontent version very likely
+        test.testFromXContent();
+    }
 }