SONAR-14882 Restore support for BOM in RulesDefinitionXmlLoader

3 files changed, 30 insertions, 1 deletions

diff --git a/sonar-plugin-api/src/main/java/org/sonar/api/server/rule/RulesDefinitionXmlLoader.java b/sonar-plugin-api/src/main/java/org/sonar/api/server/rule/RulesDefinitionXmlLoader.java
index fb5e0494766..034ee858dfb 100644
--- a/sonar-plugin-api/src/main/java/org/sonar/api/server/rule/RulesDefinitionXmlLoader.java
+++ b/sonar-plugin-api/src/main/java/org/sonar/api/server/rule/RulesDefinitionXmlLoader.java
@@ -34,6 +34,8 @@ import javax.xml.stream.XMLStreamException;
 import javax.xml.stream.events.Attribute;
 import javax.xml.stream.events.StartElement;
 import javax.xml.stream.events.XMLEvent;
+import org.apache.commons.io.ByteOrderMark;
+import org.apache.commons.io.input.BOMInputStream;
 import org.apache.commons.lang.StringUtils;
 import org.sonar.api.ce.ComputeEngineSide;
 import org.sonar.api.rule.RuleStatus;
@@ -208,7 +210,9 @@ public class RulesDefinitionXmlLoader {
    * @since 5.1
    */
   public void load(RulesDefinition.NewRepository repo, InputStream input, Charset charset) {
-    try (Reader reader = new InputStreamReader(input, charset)) {
+    try (Reader reader = new InputStreamReader(new BOMInputStream(input,
+      ByteOrderMark.UTF_8, ByteOrderMark.UTF_16LE, ByteOrderMark.UTF_16BE,
+      ByteOrderMark.UTF_32LE, ByteOrderMark.UTF_32BE), charset)) {
       load(repo, reader);
     } catch (IOException e) {
       throw new IllegalStateException("Error while reading XML rules definition for repository " + repo.key(), e);
diff --git a/sonar-plugin-api/src/test/java/org/sonar/api/server/rule/RulesDefinitionXmlLoaderTest.java b/sonar-plugin-api/src/test/java/org/sonar/api/server/rule/RulesDefinitionXmlLoaderTest.java
index 5dc8172d149..c45475af3a9 100644
--- a/sonar-plugin-api/src/test/java/org/sonar/api/server/rule/RulesDefinitionXmlLoaderTest.java
+++ b/sonar-plugin-api/src/test/java/org/sonar/api/server/rule/RulesDefinitionXmlLoaderTest.java
@@ -116,6 +116,20 @@ public class RulesDefinitionXmlLoaderTest {
   }
 
   @Test
+  public void test_utf8_encoding_with_bom() {
+    InputStream input = getClass().getResourceAsStream("RulesDefinitionXmlLoaderTest/utf8-with-bom.xml");
+    RulesDefinition.Repository repository = load(input, StandardCharsets.UTF_8.name());
+
+    assertThat(repository.rules()).hasSize(1);
+    RulesDefinition.Rule rule = repository.rules().get(0);
+    assertThat(rule.key()).isEqualTo("com.puppycrawl.tools.checkstyle.checks.naming.LocalVariableNameCheck");
+    assertThat(rule.name()).isEqualTo("M & M");
+    assertThat(rule.htmlDescription().charAt(0)).isEqualTo('\u00E9');
+    assertThat(rule.htmlDescription().charAt(1)).isEqualTo('\u00E0');
+    assertThat(rule.htmlDescription().charAt(2)).isEqualTo('\u0026');
+  }
+
+  @Test
   public void support_deprecated_format() {
     // the deprecated format uses some attributes instead of nodes
     InputStream input = getClass().getResourceAsStream("RulesDefinitionXmlLoaderTest/deprecated.xml");
diff --git a/sonar-plugin-api/src/test/resources/org/sonar/api/server/rule/RulesDefinitionXmlLoaderTest/utf8-with-bom.xml b/sonar-plugin-api/src/test/resources/org/sonar/api/server/rule/RulesDefinitionXmlLoaderTest/utf8-with-bom.xml
new file mode 100644
index 00000000000..0c0b5c88c33
--- /dev/null
+++ b/sonar-plugin-api/src/test/resources/org/sonar/api/server/rule/RulesDefinitionXmlLoaderTest/utf8-with-bom.xml
@@ -0,0 +1,11 @@
+<rules>
+  <rule>
+    <key>com.puppycrawl.tools.checkstyle.checks.naming.LocalVariableNameCheck</key>
+    <priority>BLOCKER</priority>
+    <configKey>Checker/TreeWalker/LocalVariableName</configKey>
+    <name>M &amp; M</name>
+    <description>
+      <![CDATA[éà&]]>
+    </description>
+  </rule>
+</rules>