Preserve the mark when buffering

So that we can rewind if required. Fixes #1324
jhy · Feb 15, 2020 · 62c0595 · 62c0595
1 parent 528ba55
commit 62c0595
Show file tree

Hide file tree

Showing 7 changed files with 1,048 additions and 11 deletions.
diff --git a/CHANGES b/CHANGES
@@ -13,6 +13,10 @@ jsoup changelog
   * Bugfix: in a <select> tag, a second <optgroup> would not automatically close an earlier open <optgroup>
     <https://github.com/jhy/jsoup/issues/1313>
 
+  * Bugfix: in CharacterReader when parsing an input stream, could throw a Mark Invalid exception if the reader was
+    marked, a bufferUp occurred, and then the reader was rewound.
+    <https://github.com/jhy/jsoup/issues/1324>
+
   * Removed old methods and classes that were marked deprecated in previous releases.
 
 **** Release 1.12.2 [2020-Feb-08]

diff --git a/src/main/java/org/jsoup/parser/CharacterReader.java b/src/main/java/org/jsoup/parser/CharacterReader.java
@@ -46,12 +46,18 @@ public CharacterReader(String input) {
 
     private boolean readFully; // if the underlying stream has been completely read, no value in further buffering
     private void bufferUp() {
-        if (readFully)
+        if (readFully || bufPos < bufSplitPoint)
             return;
 
-        final int pos = bufPos;
-        if (pos < bufSplitPoint)
-            return;
+        final int pos;
+        final int offset;
+        if (bufMark != -1) {
+            pos = bufMark;
+            offset = bufPos - bufMark;
+        } else {
+            pos = bufPos;
+            offset = 0;
+        }
 
         try {
             final long skipped = reader.skip(pos);
@@ -70,8 +76,9 @@ private void bufferUp() {
                 Validate.isTrue(skipped == pos); // Previously asserted that there is room in buf to skip, so this will be a WTF
                 bufLength = read;
                 readerPos += pos;
-                bufPos = 0;
-                bufMark = -1;
+                bufPos = offset;
+                if (bufMark != -1)
+                    bufMark = 0;
                 bufSplitPoint = bufLength > readAheadLimit ? readAheadLimit : bufLength;
             }
         } catch (IOException e) {

diff --git a/src/main/java/org/jsoup/parser/Token.java b/src/main/java/org/jsoup/parser/Token.java
@@ -252,7 +252,7 @@ final static class EndTag extends Tag{
 
         @Override
         public String toString() {
-            return "</" + name() + ">";
+            return "</" + (tagName != null ? tagName : "(unset)") + ">";
         }
     }
 

diff --git a/src/test/java/org/jsoup/integration/ParseTest.java b/src/test/java/org/jsoup/integration/ParseTest.java
@@ -172,6 +172,19 @@ public void testLowercaseUtf8Charset() throws IOException {
         assertEquals("UTF-8", doc.outputSettings().charset().name());
     }
 
+    @Test
+    public void testXwiki() throws IOException {
+        // https://github.com/jhy/jsoup/issues/1324
+        File in = getFile("/htmltests/xwiki-1324.html");
+        Document doc = Jsoup.parse(in, null, "https://localhost/");
+        assertEquals("XWiki Jetty HSQLDB 12.1-SNAPSHOT", doc.select("#xwikiplatformversion").text());
+
+        // was getting busted at =userdirectory, because it hit the bufferup point but the mark was then lost. so
+        // updated to preserve the mark.
+        String wantHtml = "<a class=\"list-group-item\" data-id=\"userdirectory\" href=\"/xwiki/bin/admin/XWiki/XWikiPreferences?editor=globaladmin&amp;section=userdirectory\" title=\"Customize the user directory live table.\">User Directory</a>";
+        assertEquals(wantHtml, doc.select("[data-id=userdirectory]").outerHtml());
+    }
+
     public static File getFile(String resourceName) {
         try {
             URL resource = ParseTest.class.getResource(resourceName);

diff --git a/src/test/java/org/jsoup/parser/CharacterReaderTest.java b/src/test/java/org/jsoup/parser/CharacterReaderTest.java
@@ -5,9 +5,7 @@
 import java.io.BufferedReader;
 import java.io.StringReader;
 
-import static org.junit.Assert.assertEquals;
-import static org.junit.Assert.assertFalse;
-import static org.junit.Assert.assertTrue;
+import static org.junit.Assert.*;
 
 /**
  * Test suite for character reader.
@@ -338,5 +336,4 @@ public void notEmptyAtBufferSplitPoint() {
         assertTrue(r.isEmpty());
     }
 
-
 }
diff --git a/src/test/resources/htmltests/README b/src/test/resources/htmltests/README
@@ -19,3 +19,4 @@ Sources
 * yahoo-jp.html			  http://www.yahoo.co.jp/index.html	12-Jan-2010
 * baidu-cn-home.html	  http://www.baidu.com/ 15-Jul-2010
 * nyt-article-1.html      http://www.nytimes.com/2010/07/26/business/global/26bp.html?hp
+* xwiki-1324.html         https://github.com/jhy/jsoup/issues/1324  15-Feb-2020