Check if a file contains a file link.

w3c · Nov 27, 2017 · 8f7a2b7 · 8f7a2b7
1 parent 5e33645
commit 8f7a2b7
Show file tree

Hide file tree

Showing 11 changed files with 185 additions and 33 deletions.
diff --git a/src/main/java/com/adobe/epubcheck/ctc/EntitySearch.java b/src/main/java/com/adobe/epubcheck/ctc/EntitySearch.java
@@ -20,13 +20,8 @@
 import com.adobe.epubcheck.ocf.EncryptionFilter;
 import com.adobe.epubcheck.util.EPUBVersion;
 
-class EntitySearch
+class EntitySearch extends TextSearch
 {
-  private final ZipFile zip;
-  private final Hashtable<String, EncryptionFilter> enc;
-  private final Report report;
-  private final EPUBVersion version;
-
   static final Pattern entityPattern = Pattern.compile("&([A-Za-z0-9]+)([;|\\s])");
   static final HashSet<String> legalEntities2_0;
   static final HashSet<String> legalEntities3_0;
@@ -71,32 +66,9 @@ class EntitySearch
 
   public EntitySearch(EPUBVersion version, ZipFile zip, Report report)
   {
-    this.zip = zip;
-    this.enc = new Hashtable<String, EncryptionFilter>();
-    this.report = report;
-    this.version = version;
+    super(version, zip, report);
   }
 
-  InputStream getInputStream(String name) throws
-      IOException
-  {
-    ZipEntry entry = zip.getEntry(name);
-    if (entry == null)
-    {
-      return null;
-    }
-    InputStream in = zip.getInputStream(entry);
-    EncryptionFilter filter = enc.get(name);
-    if (filter == null)
-    {
-      return in;
-    }
-    if (filter.canDecrypt())
-    {
-      return filter.decrypt(in);
-    }
-    return null;
-  }
 
   public Vector<String> Search(String entry)
   {

diff --git a/src/main/java/com/adobe/epubcheck/ctc/EpubTextContentCheck.java b/src/main/java/com/adobe/epubcheck/ctc/EpubTextContentCheck.java
@@ -1,5 +1,7 @@
 package com.adobe.epubcheck.ctc;
 
+import java.util.ArrayList;
+import java.util.List;
 import java.util.zip.ZipEntry;
 
 import com.adobe.epubcheck.api.Report;
@@ -13,12 +15,14 @@ public class EpubTextContentCheck implements DocumentValidator
 {
   private final Report report;
   private final EpubPackage epack;
-  private final EntitySearch search;
+  private final List<TextSearch> search;
 
   public EpubTextContentCheck(Report report, EpubPackage epack)
   {
     this.epack = epack;
-    this.search = new EntitySearch(epack.getVersion(), epack.getZip(), report);
+    this.search = new ArrayList<TextSearch>();
+    this.search.add(new EntitySearch(epack.getVersion(), epack.getZip(), report));
+    this.search.add(new FileLinkSearch(epack.getVersion(), epack.getZip(), report));
     this.report = report;
   }
 
@@ -40,7 +44,11 @@ public boolean validate()
           // report.message(MessageId.RSC_001, EPUBLocation.create(this.epack.getFileName()), fileToParse);
           continue;
         }
-        this.search.Search(fileToParse);
+
+        for(TextSearch ts : this.search)
+        {
+          ts.Search(fileToParse);
+        }
       }
     }
     return true;

diff --git a/src/main/java/com/adobe/epubcheck/ctc/FileLinkSearch.java b/src/main/java/com/adobe/epubcheck/ctc/FileLinkSearch.java
@@ -0,0 +1,82 @@
+package com.adobe.epubcheck.ctc;
+
+import com.adobe.epubcheck.api.EPUBLocation;
+import com.adobe.epubcheck.api.Report;
+import com.adobe.epubcheck.messages.MessageId;
+import com.adobe.epubcheck.util.EPUBVersion;
+
+import java.io.*;
+import java.util.Scanner;
+import java.util.Vector;
+import java.util.regex.Matcher;
+import java.util.regex.Pattern;
+import java.util.zip.ZipFile;
+
+public class FileLinkSearch extends TextSearch {
+    private static final Pattern fileLinkPattern = Pattern.compile("href=[\"']file://");
+
+    public FileLinkSearch(EPUBVersion version, ZipFile zip, Report report)
+    {
+        super(version, zip, report);
+    }
+
+    @Override
+    Vector<String> Search(String entry)
+    {
+        String fileName = new File(zip.getName()).getName();
+        InputStream is = null;
+        BufferedReader br = null;
+        try
+        {
+            is = getInputStream(entry);
+            br = new BufferedReader(new InputStreamReader(is));
+
+            int lineCounter = 1;
+            String line;
+            while ((line = br.readLine()) != null)
+            {
+                Matcher matcher = fileLinkPattern.matcher(line);
+                int position = 0;
+                while (matcher.find(position))
+                {
+                    int contextStart = Math.max(0, matcher.start() - 20);
+                    int contextEnd = Math.min(contextStart + 40, line.length() - 1);
+                    String context = line.substring(contextStart, contextEnd);
+
+                    report.message(MessageId.HTM_053, EPUBLocation.create(entry, lineCounter, matcher.start(), context.trim()), context.trim());
+                    position = matcher.end();
+                }
+                lineCounter++;
+            }
+        }
+        catch (FileNotFoundException e1)
+        {
+            report.message(MessageId.RSC_001, EPUBLocation.create(fileName), entry);
+        }
+        catch (IOException e1)
+        {
+            report.message(MessageId.PKG_008, EPUBLocation.create(fileName), entry);
+        }
+        catch (Exception e)
+        {
+            e.printStackTrace();
+            report.message(MessageId.RSC_005, EPUBLocation.create(entry), e.getMessage());
+        }
+        finally
+        {
+            silentlyClose(br);
+            silentlyClose(is);
+        }
+        return new Vector<String>();
+    }
+
+    private void silentlyClose(Closeable c) {
+        try
+        {
+            c.close();
+        }
+        catch (IOException ignored)
+        {
+        }
+    }
+}
diff --git a/src/main/java/com/adobe/epubcheck/ctc/TextSearch.java b/src/main/java/com/adobe/epubcheck/ctc/TextSearch.java
@@ -0,0 +1,50 @@
+package com.adobe.epubcheck.ctc;
+
+import com.adobe.epubcheck.api.Report;
+import com.adobe.epubcheck.ocf.EncryptionFilter;
+import com.adobe.epubcheck.util.EPUBVersion;
+
+import java.io.IOException;
+import java.io.InputStream;
+import java.util.Hashtable;
+import java.util.Vector;
+import java.util.zip.ZipEntry;
+import java.util.zip.ZipFile;
+
+abstract class TextSearch {
+    private final Hashtable<String, EncryptionFilter> enc;
+    final ZipFile zip;
+    final Report report;
+    final EPUBVersion version;
+
+
+    public TextSearch(EPUBVersion version, ZipFile zip, Report report)
+    {
+        this.zip = zip;
+        this.enc = new Hashtable<String, EncryptionFilter>();
+        this.report = report;
+        this.version = version;
+    }
+
+    InputStream getInputStream(String name) throws IOException
+    {
+        ZipEntry entry = zip.getEntry(name);
+        if (entry == null)
+        {
+            return null;
+        }
+        InputStream in = zip.getInputStream(entry);
+        EncryptionFilter filter = enc.get(name);
+        if (filter == null)
+        {
+            return in;
+        }
+        if (filter.canDecrypt())
+        {
+            return filter.decrypt(in);
+        }
+        return null;
+    }
+
+    abstract Vector<String> Search(String entry);
+}
diff --git a/src/main/java/com/adobe/epubcheck/messages/MessageDictionary.java b/src/main/java/com/adobe/epubcheck/messages/MessageDictionary.java
@@ -164,6 +164,7 @@ static Map<MessageId, Severity> getDefaultSeverities()
       map.put(MessageId.HTM_050, Severity.USAGE);
       map.put(MessageId.HTM_051, Severity.WARNING);
       map.put(MessageId.HTM_052, Severity.ERROR);
+      map.put(MessageId.HTM_053, Severity.INFO);
 
       // Media
       map.put(MessageId.MED_001, Severity.ERROR);

diff --git a/src/main/java/com/adobe/epubcheck/messages/MessageId.java b/src/main/java/com/adobe/epubcheck/messages/MessageId.java
@@ -127,6 +127,7 @@ public enum MessageId implements Comparable<MessageId>
   HTM_050("HTM-050"),
   HTM_051("HTM-051"),
   HTM_052("HTM-052"),
+  HTM_053("HTM_053"),
 
   // Messages associated with media (images, audio and video)
   MED_001("MED-001"),

diff --git a/src/main/resources/com/adobe/epubcheck/messages/MessageBundle.properties b/src/main/resources/com/adobe/epubcheck/messages/MessageBundle.properties
@@ -121,6 +121,7 @@ HTM_049_SUG=Add xmlns="http://www.w3.org/1999/xhtml" to the html element.
 HTM_050=Found epub:type="pagebreak" attribute in content document.
 HTM_051=Found Microdata semantic enrichments but no RDFa. EDUPUB recommends using RDFa Lite.
 HTM_052=The property 'region-based' is only allowed on nav elements in Data Navigation Documents.
+HTM_053=Found an external file link (file://) in file: '%1$s'.
 
 #media
 MED_001=Video poster must have core media image type.

diff --git a/src/test/java/com/adobe/epubcheck/api/AbstractEpubCheckTest.java b/src/test/java/com/adobe/epubcheck/api/AbstractEpubCheckTest.java
@@ -23,7 +23,9 @@
 package com.adobe.epubcheck.api;
 
 import static org.junit.Assert.assertEquals;
+import static org.junit.Assert.assertThat;
 import static org.junit.Assert.assertTrue;
+import static org.junit.matchers.JUnitMatchers.hasItems;
 
 import java.io.BufferedReader;
 import java.io.File;
@@ -54,6 +56,8 @@ public abstract class AbstractEpubCheckTest
   List<MessageId> expectedWarnings = new LinkedList<MessageId>();
   List<MessageId> expectedErrors = new LinkedList<MessageId>();
   List<MessageId> expectedFatals = new LinkedList<MessageId>();
+  List<MessageId> expectedInfos = new LinkedList<MessageId>();
+
 
   protected AbstractEpubCheckTest(String basepath)
   {
@@ -148,6 +152,9 @@ public void testValidateDocument(String fileName, String resultFile, EPUBProfile
         IdsToListOfString(testReport.getWarningIds()));
     assertEquals("The fatal error results do not match", IdsToListOfString(expectedFatals),
         IdsToListOfString(testReport.getFatalErrorIds()));
+    assertThat("The info results do not match",
+            IdsToListOfString(testReport.getInfoIds()),
+            hasItems(IdsToListOfString(expectedInfos).toArray(new String[expectedInfos.size()])));
 
     if (resultFile != null)
     {

diff --git a/src/test/java/com/adobe/epubcheck/api/Epub30CheckTest.java b/src/test/java/com/adobe/epubcheck/api/Epub30CheckTest.java
@@ -188,6 +188,13 @@ public void testValidateEPUB30Issue221()
     testValidateDocument("invalid/issue221.epub");
   }
 
+  @Test
+  public void testValidateEPUB30Issue289()
+  {
+    Collections.addAll(expectedInfos, MessageId.HTM_053);
+    testValidateDocument("valid/issue289.epub");
+  }
+
   @Test
   public void testValidateEPUB30FontObfuscation()
   {

diff --git a/src/test/java/com/adobe/epubcheck/util/ValidationReport.java b/src/test/java/com/adobe/epubcheck/util/ValidationReport.java
@@ -97,6 +97,10 @@ else if (message.getSeverity().equals(Severity.FATAL))
     {
       fatalError(PathUtil.removeWorkingDirectory(location.getPath()), location.getLine(), location.getColumn(), message.getMessage(args), message.getID());
     }
+    else if (message.getSeverity().equals(Severity.INFO))
+    {
+      info(PathUtil.removeWorkingDirectory(location.getPath()), location.getLine(), location.getColumn(), message.getMessage(args), message.getID());
+    }
   }
 
   private void error(String resource, int line, int column, String message, MessageId id)
@@ -117,6 +121,12 @@ private void warning(String resource, int line, int column, String message, Mess
     warningList.add(item);
   }
 
+  public void info(String resource, int line, int column, String message, MessageId id)
+  {
+    ItemReport item = new ItemReport(resource, line, column, fixMessage(message), id);
+    getInfoList().add(item);
+  }
+
   public String toString()
   {
     StringBuilder buffer = new StringBuilder();
@@ -233,6 +243,18 @@ public int generate()
     return 0;
   }
 
+  public List<MessageId> getInfoIds()
+  {
+    List<MessageId> result = new ArrayList<MessageId>();
+    for (ItemReport it : infoList)
+    {
+      if(it.id != null) {
+        result.add(it.id);
+      }
+    }
+    return result;
+  }
+
   public List<MessageId> getWarningIds()
   {
     List<MessageId> result = new ArrayList<MessageId>();
@@ -262,4 +284,5 @@ public List<MessageId> getFatalErrorIds()
     }
     return result;
   }
+
 }
diff --git a/src/test/resources/30/epub/valid/issue289.epub b/src/test/resources/30/epub/valid/issue289.epub