Fix title/source spans for link reference definitions with unclosed titles

CHANGELOG.md

@@ -6,6 +6,12 @@ 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]
+### Fixed
+- Fix parsing of unclosed titles for link reference definitions, fix source
+  spans in resulting paragraph (#460)
+
+
 ## [0.30.1] - 2026-10-11
 ### Changed
 - Improve MarkdownRenderer performance by appending unescaped spans in bulk (#456)
@@ -601,6 +607,7 @@ API breaking changes (caused by changes in spec):
 Initial release of commonmark-java, a port of commonmark.js with extensions
 for autolinking URLs, GitHub flavored strikethrough and tables.
 
+[Unreleased]: https://github.com/commonmark/commonmark-java/compare/commonmark-parent-0.30.1...HEAD
 [0.30.1]: https://github.com/commonmark/commonmark-java/compare/commonmark-parent-0.30.0...commonmark-parent-0.30.1
 [0.30.0]: https://github.com/commonmark/commonmark-java/compare/commonmark-parent-0.29.0...commonmark-parent-0.30.0
 [0.29.0]: https://github.com/commonmark/commonmark-java/compare/commonmark-parent-0.28.0...commonmark-parent-0.29.0

commonmark/src/main/java/org/commonmark/internal/LinkReferenceDefinitionParser.java

@@ -31,6 +31,7 @@ public class LinkReferenceDefinitionParser {
     private char titleDelimiter;
     private StringBuilder title;
     private boolean referenceValid = false;
+    private int referenceSourceSpanCount;
 
     public void parse(SourceLine line) {
         paragraphLines.add(line);
@@ -87,6 +88,11 @@ public void parse(SourceLine line) {
 
     public void addSourceSpan(SourceSpan sourceSpan) {
         sourceSpans.add(sourceSpan);
+        // Clearing paragraphLines commits the current line to a valid definition. Its span
+        // arrives after parsing, so record the boundary here.
+        if (paragraphLines.isEmpty()) {
+            referenceSourceSpanCount = sourceSpans.size();
+        }
     }
 
     /**
@@ -278,10 +284,14 @@ private void finishReference() {
         }
 
         String d = Escaping.unescapeString(destination);
-        String t = title != null ? Escaping.unescapeString(title.toString()) : null;
+        String t =
+                state != State.TITLE && title != null
+                        ? Escaping.unescapeString(title.toString())
+                        : null;
         LinkReferenceDefinition definition = new LinkReferenceDefinition(label.toString(), d, t);
-        definition.setSourceSpans(sourceSpans);
-        sourceSpans.clear();
+        definition.setSourceSpans(sourceSpans.subList(0, referenceSourceSpanCount));
+        sourceSpans.subList(0, referenceSourceSpanCount).clear();
+        referenceSourceSpanCount = 0;
         definitions.add(definition);
 
         label = null;

commonmark/src/test/java/org/commonmark/internal/LinkReferenceDefinitionParserTest.java

@@ -4,6 +4,7 @@
 
 import org.commonmark.internal.LinkReferenceDefinitionParser.State;
 import org.commonmark.node.LinkReferenceDefinition;
+import org.commonmark.node.SourceSpan;
 import org.commonmark.parser.SourceLine;
 import org.junit.jupiter.api.Test;
 
@@ -159,14 +160,64 @@ void testTitleMultiline3() {
 
     @Test
     void testTitleMultiline4() {
+        var definitionSpan = SourceSpan.of(0, 0, 0, 11);
+        var titleSpan = SourceSpan.of(1, 0, 12, 6);
+        var invalidSpan = SourceSpan.of(2, 0, 19, 4);
         parse("[foo]: /url");
+        parser.addSourceSpan(definitionSpan);
         assertThat(parser.getState()).isEqualTo(State.START_TITLE);
         parse("(title");
+        parser.addSourceSpan(titleSpan);
         assertThat(parser.getState()).isEqualTo(State.TITLE);
         parse("foo(");
+        parser.addSourceSpan(invalidSpan);
         assertThat(parser.getState()).isEqualTo(State.PARAGRAPH);
 
         assertDef(parser.getDefinitions().get(0), "foo", "/url", null);
+        assertThat(parser.getDefinitions().get(0).getSourceSpans()).containsExactly(definitionSpan);
+        assertParagraphLines("(title\nfoo(", parser);
+        assertThat(parser.getParagraphSourceSpans()).containsExactly(titleSpan, invalidSpan);
+    }
+
+    @Test
+    void testUnclosedMultilineTitleKeepsParagraphSourceSpans() {
+        var definitionSpan = SourceSpan.of(0, 0, 0, 11);
+        var paragraphSpan = SourceSpan.of(1, 0, 12, 14);
+
+        parse("[foo]: /url");
+        parser.addSourceSpan(definitionSpan);
+        parse("(partial title");
+        parser.addSourceSpan(paragraphSpan);
+
+        var definitions = parser.getDefinitions();
+
+        assertThat(definitions).hasSize(1);
+        assertDef(definitions.get(0), "foo", "/url", null);
+        assertThat(definitions.get(0).getSourceSpans()).containsExactly(definitionSpan);
+        assertParagraphLines("(partial title", parser);
+        assertThat(parser.getParagraphSourceSpans()).containsExactly(paragraphSpan);
+    }
+
+    @Test
+    void testMultilineTitleWithTrailingGarbageKeepsParagraphSourceSpans() {
+        var definitionSpan = SourceSpan.of(0, 0, 0, 11);
+        var titleSpan = SourceSpan.of(1, 0, 12, 6);
+        var invalidSpan = SourceSpan.of(2, 0, 19, 18);
+
+        parse("[foo]: /url");
+        parser.addSourceSpan(definitionSpan);
+        parse("\"title");
+        parser.addSourceSpan(titleSpan);
+        parse("continued\" garbage");
+        parser.addSourceSpan(invalidSpan);
+
+        var definitions = parser.getDefinitions();
+
+        assertThat(definitions).hasSize(1);
+        assertDef(definitions.get(0), "foo", "/url", null);
+        assertThat(definitions.get(0).getSourceSpans()).containsExactly(definitionSpan);
+        assertParagraphLines("\"title\ncontinued\" garbage", parser);
+        assertThat(parser.getParagraphSourceSpans()).containsExactly(titleSpan, invalidSpan);
     }
 
     @Test