Allow Detection of WWW Autolinks

CHANGELOG.md

@@ -6,6 +6,24 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
 This project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.html),
 with the exception that 0.x versions can break between minor versions.
 
+## Unreleased
+### Added
+- Autolink extension: Now supports configuration of different link types that
+  should be recognized and converted to links. See `AutolinkExtension#builder`
+
+  | Type    | Default? | Description                                            |
+  |---------|----------|--------------------------------------------------------|
+  | `URL`   | Yes      | URL with a protocol such as `https://example.com`      |
+  | `EMAIL` | Yes      | Email address such as `foo@example.com`                |
+  | `WWW`   | Yes      | Address beginning with `www` such as `www.example.com` |
+
+  Note that this changes the behavior of `AutolinkExtension.create()` to now also
+  include `WWW` links by default. To re-enable the previous behavior, use:
+
+  ```java
+  AutolinkExtension.builder().linkTypes(AutolinkType.URL, AutolinkType.EMAIL).build();
+  ```
+
 ## [0.26.0] - 2025-09-13
 ### Changed
 - A `LinkProcessor` using `replaceWith` now also stops outer links from being

commonmark-ext-autolink/src/main/java/org/commonmark/ext/autolink/AutolinkExtension.java

@@ -1,5 +1,8 @@
 package org.commonmark.ext.autolink;
 
+import java.util.EnumSet;
+import java.util.Set;
+
 import org.commonmark.Extension;
 import org.commonmark.ext.autolink.internal.AutolinkPostProcessor;
 import org.commonmark.parser.Parser;
@@ -18,16 +21,71 @@
  */
 public class AutolinkExtension implements Parser.ParserExtension {
 
-    private AutolinkExtension() {
+    private final Set<AutolinkType> linkTypes;
+
+    private AutolinkExtension(Builder builder) {
+        this.linkTypes = builder.linkTypes;
     }
 
+    /**
+     * @return the extension with default options
+     */
     public static Extension create() {
-        return new AutolinkExtension();
+        return builder().build();
+    }
+
+    /**
+     * @return a builder to configure the behavior of the extension.
+     */
+    public static Builder builder() {
+        return new Builder();
     }
 
     @Override
     public void extend(Parser.Builder parserBuilder) {
-        parserBuilder.postProcessor(new AutolinkPostProcessor());
+        parserBuilder.postProcessor(new AutolinkPostProcessor(linkTypes));
     }
 
+    public static class Builder {
+
+        private Set<AutolinkType> linkTypes = EnumSet.allOf(AutolinkType.class);
+
+        /**
+         * @param linkTypes the link types that should be converted. By default,
+         *                  all {@link AutolinkType}s are converted.
+         * @return {@code this}
+         */
+        public Builder linkTypes(AutolinkType... linkTypes) {
+            if (linkTypes == null) {
+                throw new NullPointerException("linkTypes must not be null");
+            }
+
+            return this.linkTypes(Set.of(linkTypes));
+        }
+
+        /**
+         * @param linkTypes the link types that should be converted. By default,
+         *                  all {@link AutolinkType}s are converted.
+         * @return {@code this}
+         */
+        public Builder linkTypes(Set<AutolinkType> linkTypes) {
+            if (linkTypes == null) {
+                throw new NullPointerException("linkTypes must not be null");
+            }
+
+            if (linkTypes.isEmpty()) {
+                throw new IllegalArgumentException("linkTypes must not be empty");
+            }
+
+            this.linkTypes = EnumSet.copyOf(linkTypes);
+            return this;
+        }
+
+        /**
+         * @return a configured extension
+         */
+        public Extension build() {
+            return new AutolinkExtension(this);
+        }
+    }
 }

commonmark-ext-autolink/src/main/java/org/commonmark/ext/autolink/AutolinkType.java

@@ -0,0 +1,19 @@
+package org.commonmark.ext.autolink;
+
+/**
+ * The types of strings that can be automatically turned into links.
+ */
+public enum AutolinkType {
+    /**
+     * URL such as {@code http://example.com}
+     */
+    URL,
+    /**
+     * Email address such as {@code foo@example.com}
+     */
+    EMAIL,
+    /**
+     * URL such as {@code www.example.com}
+     */
+    WWW
+}

commonmark-ext-autolink/src/main/java/org/commonmark/ext/autolink/internal/AutolinkPostProcessor.java

@@ -1,5 +1,6 @@
 package org.commonmark.ext.autolink.internal;
 
+import org.commonmark.ext.autolink.AutolinkType;
 import org.commonmark.node.*;
 import org.commonmark.parser.PostProcessor;
 import org.nibor.autolink.LinkExtractor;
@@ -11,9 +12,36 @@
 
 public class AutolinkPostProcessor implements PostProcessor {
 
-    private LinkExtractor linkExtractor = LinkExtractor.builder()
-            .linkTypes(EnumSet.of(LinkType.URL, LinkType.EMAIL))
-            .build();
+    private final LinkExtractor linkExtractor;
+
+    public AutolinkPostProcessor(Set<AutolinkType> linkTypes) {
+        if (linkTypes == null) {
+            throw new NullPointerException("linkTypes must not be null");
+        }
+
+        if (linkTypes.isEmpty()) {
+            throw new IllegalArgumentException("linkTypes must not be empty");
+        }
+
+        var types = EnumSet.noneOf(LinkType.class);
+        for (AutolinkType linkType : linkTypes) {
+            switch (linkType) {
+                case URL:
+                    types.add(LinkType.URL);
+                    break;
+                case EMAIL:
+                    types.add(LinkType.EMAIL);
+                    break;
+                case WWW:
+                    types.add(LinkType.WWW);
+                    break;
+            }
+        }
+
+        this.linkExtractor = LinkExtractor.builder()
+                .linkTypes(types)
+                .build();
+    }
 
     @Override
     public Node process(Node node) {
@@ -67,8 +95,13 @@ private static Text createTextNode(String literal, Span span, SourceSpan sourceS
     }
 
     private static String getDestination(LinkSpan linkSpan, String linkText) {
-        if (linkSpan.getType() == LinkType.EMAIL) {
+        var type = linkSpan.getType();
+
+        if (type == LinkType.EMAIL) {
             return "mailto:" + linkText;
+        } else if (type == LinkType.WWW) {
+            // Use http instead of https (see https://github.github.com/gfm/#extended-www-autolink)
+            return "http://" + linkText;
         } else {
             return linkText;
         }

commonmark-ext-autolink/src/test/java/org/commonmark/ext/autolink/AutolinkTest.java

@@ -19,6 +19,12 @@ public class AutolinkTest extends RenderingTestCase {
     private static final Parser PARSER = Parser.builder().extensions(EXTENSIONS).build();
     private static final HtmlRenderer RENDERER = HtmlRenderer.builder().extensions(EXTENSIONS).build();
 
+    private static final Set<Extension> NO_WWW_EXTENSIONS = Set.of(AutolinkExtension.builder()
+        .linkTypes(AutolinkType.URL, AutolinkType.EMAIL)
+        .build());
+    private static final Parser NO_WWW_PARSER = Parser.builder().extensions(NO_WWW_EXTENSIONS).build();
+    private static final HtmlRenderer NO_WWW_RENDERER = HtmlRenderer.builder().extensions(NO_WWW_EXTENSIONS).build();
+
     @Test
     public void oneTextNode() {
         assertRendering("foo http://one.org/ bar http://two.org/",
@@ -57,6 +63,18 @@ public void dontLinkTextWithinLinks() {
                 "<p><a href=\"http://example.com\">http://example.com</a></p>\n");
     }
 
+    @Test
+    public void wwwLinks() {
+        assertRendering("www.example.com",
+                "<p><a href=\"http://www.example.com\">www.example.com</a></p>\n");
+    }
+
+    @Test
+    public void noWwwLinks() {
+        String html = NO_WWW_RENDERER.render(NO_WWW_PARSER.parse("www.example.com"));
+        assertThat(html).isEqualTo("<p>www.example.com</p>\n");
+    }
+
     @Test
     public void sourceSpans() {
         Parser parser = Parser.builder()