extract HttpHeadersUtil::parseCharacterEncoding and parseAcceptCharset (#11670)

sdelamo · web-flow · commit 19da2da06f7b · 2025-03-21T16:13:17.000+01:00
* extract HttpHeadersUtil::parseCharacterEncoding and parseAcceptCharset

To allow the usage of those methods in runtimes where no HttpRequest or HttpMessage is available

* remove unnecessary class qualifier

* fix test

* deafult charset utf-8
diff --git a/http/src/main/java/io/micronaut/http/HttpHeaders.java b/http/src/main/java/io/micronaut/http/HttpHeaders.java
@@ -21,7 +21,6 @@
 import jakarta.annotation.Nullable;
 
 import java.nio.charset.Charset;
-import java.nio.charset.StandardCharsets;
 import java.time.LocalDateTime;
 import java.time.ZoneId;
 import java.time.ZonedDateTime;
@@ -744,17 +743,7 @@ default Charset acceptCharset() {
      */
     default Optional<Charset> findAcceptCharset() {
         return findFirst(HttpHeaders.ACCEPT_CHARSET)
-            .map(text -> {
-                text = HttpHeadersUtil.splitAcceptHeader(text);
-                if (text != null) {
-                    try {
-                        return Charset.forName(text);
-                    } catch (Exception ignored) {
-                    }
-                }
-                // default to UTF-8
-                return StandardCharsets.UTF_8;
-            });
+            .map(HttpHeadersUtil::parseAcceptCharset);
     }
 
     /**
diff --git a/http/src/main/java/io/micronaut/http/util/HttpHeadersUtil.java b/http/src/main/java/io/micronaut/http/util/HttpHeadersUtil.java
@@ -20,9 +20,14 @@
 import io.micronaut.core.annotation.Nullable;
 import io.micronaut.core.util.SupplierUtil;
 import io.micronaut.http.HttpHeaders;
+import io.micronaut.http.MediaType;
 import org.slf4j.Logger;
 
+import java.nio.charset.Charset;
+import java.nio.charset.StandardCharsets;
+import java.nio.charset.UnsupportedCharsetException;
 import java.util.List;
+import java.util.Optional;
 import java.util.Set;
 import java.util.function.Function;
 import java.util.function.Supplier;
@@ -34,6 +39,7 @@
  * @since 3.8.0
  */
 public final class HttpHeadersUtil {
+    private final static Charset DEFAULT_CHARSET = StandardCharsets.UTF_8;
     private static final Supplier<Pattern> HEADER_MASK_PATTERNS = SupplierUtil.memoized(() ->
         Pattern.compile(".*(password|cred|cert|key|secret|token|auth|signat).*", Pattern.CASE_INSENSITIVE)
     );
@@ -135,4 +141,66 @@ public static String splitAcceptHeader(@NonNull String text) {
         }
         return text;
     }
+
+    /**
+     * Resolve the {@link Charset} to use for request identified by the Content-Type HTTP Header value and the Accept-Charset HTTP Header value.
+     *
+     * @param contentTypeHeaderValue Content-Type HTTP Header Value
+     * @param acceptCharsetHeaderValue Accept-Charset HTTP Header Value
+     * @return A {@link Charset}
+     * @since 4.8.8
+     */
+    @NonNull
+    public static Charset parseCharacterEncoding(@Nullable String contentTypeHeaderValue, @Nullable String acceptCharsetHeaderValue) {
+        MediaType contentType = contentTypeHeaderValue == null ? null : MediaType.of(contentTypeHeaderValue);
+        Charset charset = acceptCharsetHeaderValue != null ? parseAcceptCharset(acceptCharsetHeaderValue) : StandardCharsets.UTF_8;
+        return parseCharacterEncoding(contentType, charset);
+    }
+
+    /**
+     * Resolve the {@link Charset} to use for the request.
+     *
+     * @param contentType ContenType
+     * @param acceptCharset Accept Charset
+     * @return An {@link Optional} of {@link Charset}
+     * @since 4.8.8
+     */
+    @NonNull
+    public static Charset parseCharacterEncoding(@Nullable MediaType contentType,
+                                                 @NonNull Charset acceptCharset) {
+        try {
+
+            if (contentType != null) {
+                String charset = contentType.getParametersMap().get(MediaType.CHARSET_PARAMETER);
+                if (charset != null) {
+                    try {
+                        return Charset.forName(charset);
+                    } catch (Exception e) {
+                        return DEFAULT_CHARSET;
+                    }
+                }
+            }
+        } catch (UnsupportedCharsetException e) {
+            return DEFAULT_CHARSET;
+        }
+        return acceptCharset;
+    }
+
+    /**
+     *
+     * @param acceptCharsetHeaderValue Accept-Charset HeaderValue
+     * @return Accept Charset
+     * @since 4.8.8
+     */
+    @NonNull
+    public static Charset parseAcceptCharset(@NonNull String acceptCharsetHeaderValue) {
+        String text = HttpHeadersUtil.splitAcceptHeader(acceptCharsetHeaderValue);
+        if (text != null) {
+            try {
+                return Charset.forName(text);
+            } catch (Exception ignored) {
+            }
+        }
+        return DEFAULT_CHARSET;
+    }
 }
diff --git a/http/src/main/java/io/micronaut/http/util/HttpUtil.java b/http/src/main/java/io/micronaut/http/util/HttpUtil.java
@@ -32,6 +32,7 @@
  * @since 1.0
  */
 public class HttpUtil {
+    private static final Charset DEFAULT_CHARSET = StandardCharsets.UTF_8;
 
     /**
      * Return whether the given request features {@link MediaType#APPLICATION_FORM_URLENCODED} or
@@ -86,22 +87,8 @@ public static Optional<Charset> resolveCharset(@NonNull HttpMessage<?> request)
     @SuppressWarnings("Duplicates")
     @NonNull
     public static Charset getCharset(@NonNull HttpMessage<?> request) {
-        try {
-            MediaType contentType = request.getContentType().orElse(null);
-            if (contentType != null) {
-                String charset = contentType.getParametersMap().get(MediaType.CHARSET_PARAMETER);
-                if (charset != null) {
-                    try {
-                        return Charset.forName(charset);
-                    } catch (Exception e) {
-                        // unsupported charset, default to UTF-8
-                        return Charset.defaultCharset();
-                    }
-                }
-            }
-        } catch (UnsupportedCharsetException e) {
-            return StandardCharsets.UTF_8;
-        }
-        return request.getHeaders().findAcceptCharset().orElse(StandardCharsets.UTF_8);
+        MediaType contentType = request.getContentType().orElse(null);
+        return HttpHeadersUtil.parseCharacterEncoding(contentType,
+            request.getHeaders() != null ? request.getHeaders().findAcceptCharset().orElse(DEFAULT_CHARSET) : DEFAULT_CHARSET);
     }
 }
diff --git a/http/src/test/groovy/io/micronaut/http/util/HttpHeadersUtilSpec.groovy b/http/src/test/groovy/io/micronaut/http/util/HttpHeadersUtilSpec.groovy
@@ -6,8 +6,11 @@ import ch.qos.logback.core.AppenderBase
 import io.micronaut.http.HttpHeaders
 import org.slf4j.Logger
 import org.slf4j.LoggerFactory
+import spock.lang.See
 import spock.lang.Specification
 
+import java.nio.charset.Charset
+import java.nio.charset.StandardCharsets
 import java.util.concurrent.BlockingQueue
 import java.util.concurrent.LinkedBlockingQueue
 
@@ -78,6 +81,40 @@ class HttpHeadersUtilSpec extends Specification {
         "*"                                            | null
     }
 
+    @See("https://udn.realityripple.com/docs/Web/HTTP/Headers/Accept-Charset")
+    void "acceptCharset"(String headerValue, Charset expectedCharset) {
+        when:
+        Charset charset = HttpHeadersUtil.parseAcceptCharset(headerValue)
+
+        then:
+        charset == expectedCharset
+
+        where:
+        headerValue                        || expectedCharset
+        'iso-8859-1'                       || StandardCharsets.ISO_8859_1
+        'utf-8, iso-8859-1;q=0.5, *;q=0.1' || StandardCharsets.UTF_8
+        'utf-8, iso-8859-1;q=0.5'          || StandardCharsets.UTF_8
+
+    }
+
+    void "parse character encoding based on the Content-Type and Accept-Charset Header values"(String contentTypeHeaderValue, String acceptCharsetHeaderValue , Charset expectedCharset) {
+        when:
+        Charset charset = HttpHeadersUtil.parseCharacterEncoding(contentTypeHeaderValue, acceptCharsetHeaderValue)
+
+        then:
+        charset == expectedCharset
+
+        where:
+        contentTypeHeaderValue            | acceptCharsetHeaderValue  || expectedCharset
+        null                              | 'iso-8859-1'              || StandardCharsets.ISO_8859_1
+        null                              | 'utf-8, iso-8859-1;q=0.5, *;q=0.1' || StandardCharsets.UTF_8
+        null                              | 'utf-8, iso-8859-1;q=0.5' || StandardCharsets.UTF_8
+        'application/json; charset=utf-8' | 'iso-8859-1'              || StandardCharsets.UTF_8
+        'application/json'                | 'iso-8859-1'              || StandardCharsets.ISO_8859_1
+        null                              | null                      || StandardCharsets.UTF_8
+        'application/json; charset=bogus' | 'iso-8859-1'              || StandardCharsets.UTF_8
+    }
+
     static class MemoryAppender extends AppenderBase<ILoggingEvent> {
         final BlockingQueue<String> events = new LinkedBlockingQueue<>()
 
