get finally rid of commons codec

Author: rbri

pom.xml

@@ -36,7 +36,6 @@
         <commons-lang3.version>3.17.0</commons-lang3.version>
         <commons-io.version>2.18.0</commons-io.version>
         <commons-logging.version>1.3.5</commons-logging.version>
-        <commons-codec.version>1.18.0</commons-codec.version>
         <brotli.version>0.1.2</brotli.version>
 
 
@@ -1216,10 +1215,6 @@
             <artifactId>httpmime</artifactId>
             <version>${httpcomponents.version}</version>
             <exclusions>
-                <exclusion>
-                    <groupId>commons-codec</groupId>
-                    <artifactId>commons-codec</artifactId>
-                </exclusion>
                 <exclusion>
                     <groupId>commons-logging</groupId>
                     <artifactId>commons-logging</artifactId>
@@ -1279,11 +1274,6 @@
             <artifactId>commons-logging</artifactId>
             <version>${commons-logging.version}</version>
         </dependency>
-        <dependency>
-            <groupId>commons-codec</groupId>
-            <artifactId>commons-codec</artifactId>
-            <version>${commons-codec.version}</version>
-        </dependency>
         <dependency>
             <groupId>org.brotli</groupId>
             <artifactId>dec</artifactId>

src/main/java/org/htmlunit/javascript/host/WindowOrWorkerGlobalScopeMixin.java

@@ -87,8 +87,17 @@ public static String btoa(final String stringToEncode, final HtmlUnitScriptable
                         org.htmlunit.javascript.host.dom.DOMException.INVALID_CHARACTER_ERR);
             }
         }
+
         final byte[] bytes = stringToEncode.getBytes(StandardCharsets.ISO_8859_1);
-        return new String(Base64.getEncoder().encode(bytes), StandardCharsets.UTF_8);
+        try {
+            return new String(Base64.getEncoder().encode(bytes), StandardCharsets.UTF_8);
+        }
+        catch (final IllegalArgumentException e) {
+            throw JavaScriptEngine.asJavaScriptException(
+                    scriptable,
+                    "Failed to execute btoa(): " + e.getMessage(),
+                    org.htmlunit.javascript.host.dom.DOMException.INVALID_CHARACTER_ERR);
+        }
     }
 
     /**

src/main/java/org/htmlunit/protocol/data/DataUrlDecoder.java

@@ -22,8 +22,8 @@
 import java.nio.charset.Charset;
 import java.nio.charset.IllegalCharsetNameException;
 import java.nio.charset.UnsupportedCharsetException;
+import java.util.Base64;
 
-import org.apache.commons.codec.binary.Base64;
 import org.apache.commons.lang3.StringUtils;
 import org.htmlunit.util.MimeType;
 import org.htmlunit.util.UrlUtils;
@@ -90,10 +90,9 @@ public static DataUrlDecoder decodeDataURL(final String url) throws UnsupportedE
 
         try {
             byte[] data = url.substring(comma + 1).getBytes(charset);
-            data = UrlUtils.decodeDataUrl(data);
+            data = UrlUtils.decodeDataUrl(data, base64);
             if (base64) {
-                // the commons codec decoder skip's invalid chars
-                data = Base64.decodeBase64(data);
+                data = Base64.getDecoder().decode(data);
             }
             return new DataUrlDecoder(data, mediaType, charset);
         }

src/main/java/org/htmlunit/util/UrlUtils.java

@@ -1384,28 +1384,47 @@ public static URL removeRedundantPort(final URL url) throws MalformedURLExceptio
      * @param bytes array of URL safe characters
      * @return array of original bytes
      * @throws IllegalArgumentException in case of error
+     *
+     * @deprecated as of version 4.11.0; use {@link #decodeDataUrl(byte[], boolean)} instead
      */
+    @Deprecated
     public static byte[] decodeDataUrl(final byte[] bytes) throws IllegalArgumentException  {
+        return decodeDataUrl(bytes, false);
+    }
+
+    /**
+     * Decodes an array of URL safe 7-bit characters into an array of original bytes.
+     * Escaped characters are converted back to their original representation.
+     * @param bytes array of URL safe characters
+     * @param removeWhitespace if true don't add whitespace chars to the output
+     * @return array of original bytes
+     * @throws IllegalArgumentException in case of error
+     */
+    public static byte[] decodeDataUrl(final byte[] bytes, final boolean removeWhitespace)
+                            throws IllegalArgumentException  {
         // adapted from apache commons codec
         if (bytes == null) {
             return null;
         }
         final ByteArrayOutputStream buffer = new ByteArrayOutputStream();
         for (int i = 0; i < bytes.length; i++) {
-            final int b = bytes[i];
+            int b = bytes[i];
             if (b == '%') {
                 try {
                     final int u = digit16(bytes[++i]);
                     final int l = digit16(bytes[++i]);
-                    buffer.write((char) ((u << 4) + l));
+                    b = (u << 4) + l;
                 }
                 catch (final ArrayIndexOutOfBoundsException e) {
                     throw new IllegalArgumentException("Invalid URL encoding: ", e);
                 }
             }
-            else {
-                buffer.write(b);
+            if (removeWhitespace
+                    && (b == 9 || b == 10 || b == 12 || b == 13 || b == 32)) {
+                continue;
             }
+
+            buffer.write(b);
         }
         return buffer.toByteArray();
     }

src/test/java/org/htmlunit/ExternalTest.java

@@ -327,12 +327,6 @@ private static boolean isIgnored(@SuppressWarnings("unused") final String groupI
             return true;
         }
 
-        if ("commons-codec".equals(groupId)
-                && "commons-codec".equals(artifactId)
-                && "20041127.091804".equals(version)) {
-            return true;
-        }
-
         // version > 3.12.0 does not work with our site.xml and also not with a refactored one
         if ("maven-site-plugin".equals(artifactId)
                 && (version.startsWith("3.12.1") || version.startsWith("3.20.") || version.startsWith("3.21."))) {

src/test/java/org/htmlunit/archunit/ArchitectureTest.java

@@ -90,17 +90,6 @@ public class ArchitectureTest {
             .and().resideOutsideOfPackage("org.htmlunit.jetty..")
         .should().dependOnClassesThat().resideInAnyPackage("java.awt..");
 
-    /**
-     * Do not use org.apache.commons.codec.binary.Base64 - use the jdk instead.
-     */
-    @ArchTest
-    public static final ArchRule jdkBase64Rule = noClasses()
-        .that()
-            .resideOutsideOfPackage("org.htmlunit.jetty..")
-            .and().doNotHaveFullyQualifiedName("org.htmlunit.protocol.data.DataUrlDecoder")
-        .should().dependOnClassesThat().haveFullyQualifiedName("org.apache.commons.codec.binary.Base64");
-
-
     /**
      * JsxClasses are always in the javascript package.
      */

src/test/java/org/htmlunit/javascript/host/Window2Test.java

@@ -242,6 +242,33 @@ public void atobWhitespace() throws Exception {
         loadPageVerifyTitle2(html);
     }
 
+    /**
+     * @throws Exception if the test fails
+     */
+    @Test
+    @Alerts({"SGVsbG8gV29ybGQh", "InvalidCharacterError/DOMException",
+             "InvalidCharacterError/DOMException", "InvalidCharacterError/DOMException"})
+    public void atobNbsp() throws Exception {
+        final String html
+            = "<html><head></head><body>\n"
+            + "<script>\n"
+            + LOG_TITLE_FUNCTION
+            + "  var data = window.btoa('Hello World!');\n"
+            + "  log(data);\n"
+            + "  try {\n"
+            + "    log(window.atob('\\xA0' + data));\n"
+            + "  } catch(e) { logEx(e) }\n"
+            + "  try {\n"
+            + "    log(window.atob(data + '\\xA0'));\n"
+            + "  } catch(e) { logEx(e) }\n"
+            + "  try {\n"
+            + "    log(window.atob(data.substr(0, 2) + '\\xA0' + data.substr(2)));\n"
+            + "  } catch(e) { logEx(e) }\n"
+            + "</script>\n"
+            + "</body></html>";
+        loadPageVerifyTitle2(html);
+    }
+
     /**
      * @throws Exception if the test fails
      */