Clone Token.Characters into PendingTableCharacters

Keeps the source start/end tracking.

Fixes #1927

Author: jhy

CHANGES

@@ -24,6 +24,9 @@ Release 1.16.1 [PENDING]
     again, causing errors when fetched.
     <https://github.com/jhy/jsoup/issues/1902>
 
+  * Bugfix: when tracking input source positions, text in tables that was fostered had invalid positions.
+    <https://github.com/jhy/jsoup/issues/1927>
+
   * Bugfix: If the Document.OutputSettings class was initialized, and then Entities.escape(String) called, an NPE may be
     thrown due to a class loading circular dependency.
     <https://github.com/jhy/jsoup/issues/1910>

src/main/java/org/jsoup/parser/HtmlTreeBuilder.java

@@ -53,7 +53,7 @@ public class HtmlTreeBuilder extends TreeBuilder {
     private @Nullable Element contextElement; // fragment parse context -- could be null even if fragment parsing
     private ArrayList<Element> formattingElements; // active (open) formatting elements
     private ArrayList<HtmlTreeBuilderState> tmplInsertMode; // stack of Template Insertion modes
-    private List<String> pendingTableCharacters; // chars in table to be shifted out
+    private List<Token.Character> pendingTableCharacters; // chars in table to be shifted out
     private Token.EndTag emptyEnd; // reused empty end tag
 
     private boolean framesetOk; // if ok to go into frameset
@@ -676,14 +676,20 @@ void setFormElement(FormElement formElement) {
         this.formElement = formElement;
     }
 
-    void newPendingTableCharacters() {
+    void resetPendingTableCharacters() {
         pendingTableCharacters = new ArrayList<>();
     }
 
-    List<String> getPendingTableCharacters() {
+    List<Token.Character> getPendingTableCharacters() {
         return pendingTableCharacters;
     }
 
+    void addPendingTableCharacters(Token.Character c) {
+        // make a clone of the token to maintain its state (as Tokens are otherwise reset)
+        Token.Character clone = c.clone();
+        pendingTableCharacters.add(clone);
+    }
+
     /**
      13.2.6.3 Closing elements that have implied end tags
      When the steps below require the UA to generate implied end tags, then, while the current node is a dd element, a dt element, an li element, an optgroup element, an option element, a p element, an rb element, an rp element, an rt element, or an rtc element, the UA must pop the current node off the stack of open elements.

src/main/java/org/jsoup/parser/HtmlTreeBuilderState.java

@@ -6,7 +6,6 @@
 import org.jsoup.nodes.Document;
 import org.jsoup.nodes.DocumentType;
 import org.jsoup.nodes.Element;
-import org.jsoup.nodes.Node;
 
 import java.util.ArrayList;
 
@@ -995,7 +994,7 @@ boolean process(Token t, HtmlTreeBuilder tb) {
     InTable {
         boolean process(Token t, HtmlTreeBuilder tb) {
             if (t.isCharacter() && inSorted(tb.currentElement().normalName(), InTableFoster)) {
-                tb.newPendingTableCharacters();
+                tb.resetPendingTableCharacters();
                 tb.markInsertionMode();
                 tb.transition(InTableText);
                 return tb.process(t);
@@ -1106,25 +1105,25 @@ boolean process(Token t, HtmlTreeBuilder tb) {
                     tb.error(this);
                     return false;
                 } else {
-                    tb.getPendingTableCharacters().add(c.getData());
+                    tb.addPendingTableCharacters(c);
                 }
-            } else {// todo - don't really like the way these table character data lists are built
+            } else {
                 if (tb.getPendingTableCharacters().size() > 0) {
-                    for (String character : tb.getPendingTableCharacters()) {
-                        if (!isWhitespace(character)) {
+                    for (Token.Character c : tb.getPendingTableCharacters()) {
+                        if (!isWhitespace(c)) {
                             // InTable anything else section:
                             tb.error(this);
                             if (inSorted(tb.currentElement().normalName(), InTableFoster)) {
                                 tb.setFosterInserts(true);
-                                tb.process(new Token.Character().data(character), InBody);
+                                tb.process(c, InBody);
                                 tb.setFosterInserts(false);
                             } else {
-                                tb.process(new Token.Character().data(character), InBody);
+                                tb.process(c, InBody);
                             }
                         } else
-                            tb.insert(new Token.Character().data(character));
+                            tb.insert(c);
                     }
-                    tb.newPendingTableCharacters();
+                    tb.resetPendingTableCharacters();
                 }
                 tb.transition(tb.originalState());
                 return tb.process(t);
@@ -1759,10 +1758,6 @@ private static boolean isWhitespace(Token t) {
         return false;
     }
 
-    private static boolean isWhitespace(String data) {
-        return StringUtil.isBlank(data);
-    }
-
     private static void handleRcData(Token.StartTag startTag, HtmlTreeBuilder tb) {
         tb.tokeniser.transition(TokeniserState.Rcdata);
         tb.markInsertionMode();

src/main/java/org/jsoup/parser/Token.java

@@ -382,7 +382,7 @@ public String toString() {
         }
     }
 
-    static class Character extends Token {
+    static class Character extends Token implements Cloneable {
         private String data;
 
         Character() {
@@ -410,6 +410,14 @@ String getData() {
         public String toString() {
             return getData();
         }
+
+        @Override protected Token.Character clone() {
+            try {
+                return (Token.Character) super.clone();
+            } catch (CloneNotSupportedException e) {
+                throw new RuntimeException(e);
+            }
+        }
     }
 
     final static class CData extends Character {

src/test/java/org/jsoup/nodes/PositionTest.java

@@ -3,9 +3,12 @@
 import org.jsoup.Jsoup;
 import org.jsoup.integration.servlets.FileServlet;
 import org.jsoup.parser.Parser;
+import org.jsoup.select.NodeTraversor;
 import org.junit.jupiter.api.Test;
 
 import java.io.IOException;
+import java.util.ArrayList;
+import java.util.List;
 
 import static org.junit.jupiter.api.Assertions.*;
 
@@ -169,4 +172,23 @@ class PositionTest {
         assertEquals("17,5:779-17,12:786", item.endSourceRange().toString());
     }
 
+    @Test void tracksTableMovedText() {
+        String html = "<table>foo<tr>bar<td>baz</td>qux</tr>coo</table>";
+        Document doc = Jsoup.parse(html, TrackingParser);
+
+        List<TextNode> textNodes = new ArrayList<>();
+        NodeTraversor.traverse((Node node, int depth) -> {
+            if (node instanceof TextNode) {
+                textNodes.add((TextNode) node);
+            }
+        }, doc);
+
+        assertEquals(5, textNodes.size());
+        assertEquals("1,8:7-1,11:10", textNodes.get(0).sourceRange().toString());
+        assertEquals("1,15:14-1,18:17", textNodes.get(1).sourceRange().toString());
+        assertEquals("1,22:21-1,25:24", textNodes.get(2).sourceRange().toString());
+        assertEquals("1,30:29-1,33:32", textNodes.get(3).sourceRange().toString());
+        assertEquals("1,38:37-1,41:40", textNodes.get(4).sourceRange().toString());
+    }
+
 }