Upgrade PDFBox API to v2.0.21 (#404)

TomRoush · May 16, 2022 · 3c6c526 · 3c6c526
1 parent ad6b957
commit 3c6c526
Show file tree

Hide file tree

Showing 64 changed files with 2,072 additions and 775 deletions.
diff --git a/library/src/androidTest/java/com/tom_roush/pdfbox/pdmodel/font/PDFontTest.java b/library/src/androidTest/java/com/tom_roush/pdfbox/pdmodel/font/PDFontTest.java
@@ -18,6 +18,7 @@
 package com.tom_roush.pdfbox.pdmodel.font;
 
 import android.content.Context;
+import android.util.Log;
 
 import androidx.test.platform.app.InstrumentationRegistry;
 
@@ -28,10 +29,15 @@
 import java.io.IOException;
 import java.io.InputStream;
 import java.io.OutputStream;
+import java.net.URI;
 import java.net.URISyntaxException;
+import java.util.ArrayList;
+import java.util.List;
 
 import com.tom_roush.fontbox.ttf.TTFParser;
+import com.tom_roush.fontbox.ttf.TrueTypeCollection;
 import com.tom_roush.fontbox.ttf.TrueTypeFont;
+import com.tom_roush.fontbox.util.autodetect.FontFileFinder;
 import com.tom_roush.pdfbox.android.PDFBoxResourceLoader;
 import com.tom_roush.pdfbox.android.TestResourceGenerator;
 import com.tom_roush.pdfbox.cos.COSName;
@@ -205,6 +211,53 @@ public void testPDFox4318() throws IOException
         }
     }
 
+    @Test
+    public void testFullEmbeddingTTC() throws IOException
+    {
+        FontFileFinder fff = new FontFileFinder();
+        TrueTypeCollection ttc = null;
+        for (URI uri : fff.find())
+        {
+            if (uri.getPath().endsWith(".ttc"))
+            {
+                File file = new File(uri);
+                 Log.i("PdfBox-Android", "TrueType collection file: " + file);
+                ttc = new TrueTypeCollection(file);
+                break;
+            }
+        }
+        if (ttc == null)
+        {
+             Log.i("PdfBox-Android", "testFullEmbeddingTTC skipped, no .ttc files available");
+            return;
+        }
+
+        final List<String> names = new ArrayList<String>();
+        ttc.processAllFonts(new TrueTypeCollection.TrueTypeFontProcessor()
+        {
+            @Override
+            public void process(TrueTypeFont ttf) throws IOException
+            {
+                 Log.i("PdfBox-Android", "TrueType font in collection: " + ttf.getName());
+                names.add(ttf.getName());
+            }
+        });
+
+        TrueTypeFont ttf = ttc.getFontByName(names.get(0)); // take the first one
+         Log.i("PdfBox-Android", "TrueType font used for test: " + ttf.getName());
+
+        try
+        {
+            PDType0Font.load(new PDDocument(), ttf, false);
+        }
+        catch (IOException ex)
+        {
+            Assert.assertEquals("Full embedding of TrueType font collections not supported", ex.getMessage());
+            return;
+        }
+        Assert.fail("should have thrown IOException");
+    }
+
     private void testPDFBox3826checkFonts(byte[] byteArray, File fontFile) throws IOException
     {
         PDDocument doc = PDDocument.load(byteArray);

diff --git a/...interactive/form/MultilineFieldsTest.java → ...m/MultilineFieldsInstrumentationTest.java b/...interactive/form/MultilineFieldsTest.java → ...m/MultilineFieldsInstrumentationTest.java
@@ -24,24 +24,23 @@
 import java.io.File;
 import java.io.IOException;
 
+import com.tom_roush.pdfbox.android.PDFBoxResourceLoader;
 import com.tom_roush.pdfbox.pdmodel.PDDocument;
 import com.tom_roush.pdfbox.rendering.TestRendering;
-import com.tom_roush.pdfbox.android.PDFBoxResourceLoader;
 
 import org.junit.After;
 import org.junit.Before;
 import org.junit.Test;
 
-public class MultilineFieldsTest
+public class MultilineFieldsInstrumentationTest
 {
     private static File OUT_DIR;
     private static final String IN_DIR = "pdfbox/com/tom_roush/pdfbox/pdmodel/interactive/form";
     private static final String NAME_OF_PDF = "MultilineFields.pdf";
-    private static final String TEST_VALUE =
-        "Lorem ipsum dolor sit amet, consetetur sadipscing elitr, " +
-            "sed diam nonumy eirmod tempor invidunt ut labore et dolore magna aliquyam";
+    private static final String TEST_VALUE = "Lorem ipsum dolor sit amet, consetetur sadipscing elitr, " +
+        "sed diam nonumy eirmod tempor invidunt ut labore et dolore magna aliquyam";
 
-    Context testContext;
+    private Context testContext;
 
     private PDDocument document;
     private PDAcroForm acroForm;
@@ -51,7 +50,6 @@ public void setUp() throws IOException
     {
         testContext = InstrumentationRegistry.getInstrumentation().getContext();
         PDFBoxResourceLoader.init(testContext);
-        System.out.println("Working Directory = " + System.getProperty("user.dir"));
         document = PDDocument.load(testContext.getAssets().open(IN_DIR + "/" + NAME_OF_PDF));
         acroForm = document.getDocumentCatalog().getAcroForm();
         OUT_DIR = new File(testContext.getCacheDir(), "pdfbox-test-output");
@@ -110,4 +108,5 @@ public void tearDown() throws IOException
     {
         document.close();
     }
+
 }
diff --git a/...androidTest/java/com/tom_roush/pdfbox/pdmodel/interactive/form/PDAcroFormFlattenTest.java b/...androidTest/java/com/tom_roush/pdfbox/pdmodel/interactive/form/PDAcroFormFlattenTest.java
@@ -284,6 +284,20 @@ public void testFlattenPDFBox4788() throws IOException
       flattenAndCompare(sourceUrl, targetFileName);
    }
 
+   /**
+    * PDFBOX-4889: appearance streams with empty /BBox.
+    *
+    * @throws IOException
+    */
+   @Test
+   public void testFlattenPDFBox4889() throws IOException
+   {
+      String sourceUrl = "https://issues.apache.org/jira/secure/attachment/13005793/f1040sb%20test.pdf";
+      String targetFileName = "PDFBOX-4889.pdf";
+
+      flattenAndCompare(sourceUrl, targetFileName);
+   }
+
    /*
     * Flatten and compare with generated image samples.
     */

diff --git a/library/src/main/java/com/tom_roush/fontbox/cmap/CMapParser.java b/library/src/main/java/com/tom_roush/fontbox/cmap/CMapParser.java
@@ -16,6 +16,7 @@
  */
 package com.tom_roush.fontbox.cmap;
 
+import java.io.BufferedInputStream;
 import java.io.File;
 import java.io.FileInputStream;
 import java.io.IOException;
@@ -122,27 +123,27 @@ public CMap parse(InputStream input) throws IOException
 
                 if (previousToken != null)
                 {
-                    if (op.op.equals("usecmap"))
+                    if (op.op.equals("usecmap") && previousToken instanceof LiteralName)
                     {
                         parseUsecmap((LiteralName) previousToken, result);
                     }
-                    else if (op.op.equals("begincodespacerange"))
+                    else if (op.op.equals("begincodespacerange") && previousToken instanceof Number)
                     {
                         parseBegincodespacerange((Number) previousToken, cmapStream, result);
                     }
-                    else if (op.op.equals("beginbfchar"))
+                    else if (op.op.equals("beginbfchar") && previousToken instanceof Number)
                     {
                         parseBeginbfchar((Number) previousToken, cmapStream, result);
                     }
-                    else if (op.op.equals("beginbfrange"))
+                    else if (op.op.equals("beginbfrange") && previousToken instanceof Number)
                     {
                         parseBeginbfrange((Number) previousToken, cmapStream, result);
                     }
-                    else if (op.op.equals("begincidchar"))
+                    else if (op.op.equals("begincidchar") && previousToken instanceof Number)
                     {
                         parseBegincidchar((Number) previousToken, cmapStream, result);
                     }
-                    else if (op.op.equals("begincidrange"))
+                    else if (op.op.equals("begincidrange") && previousToken instanceof Integer)
                     {
                         parseBegincidrange((Integer) previousToken, cmapStream, result);
                     }
@@ -452,12 +453,7 @@ protected InputStream getExternalCMap(String name) throws IOException
             return PDFBoxResourceLoader.getStream("com/tom_roush/fontbox/resources/cmap/" + name);
         }
 
-        InputStream is = getClass().getResourceAsStream("/com/tom_roush/fontbox/resources/cmap/" + name);
-        if (is == null)
-        {
-            throw new IOException("Error: Could not find referenced cmap stream " + name);
-        }
-        return is;
+        return new BufferedInputStream(getClass().getResourceAsStream("/com/tom_roush/fontbox/resources/cmap/" + name));
     }
 
     private Object parseNextToken(PushbackInputStream is) throws IOException

diff --git a/library/src/main/java/com/tom_roush/fontbox/cmap/CodespaceRange.java b/library/src/main/java/com/tom_roush/fontbox/cmap/CodespaceRange.java
@@ -37,19 +37,26 @@ public class CodespaceRange
      * &lt;8140&gt; to &lt;9FFC&gt; defines a rectangular range. The high byte has to be within 0x81 and 0x9F and the
      * low byte has to be within 0x40 and 0xFC
      *
+     * @param startBytes
+     * @param endBytes
      */
     public CodespaceRange(byte[] startBytes, byte[] endBytes)
     {
-        if (startBytes.length != endBytes.length)
+        byte[] correctedStartBytes = startBytes;
+        if (startBytes.length != endBytes.length && startBytes.length == 1 && startBytes[0] == 0)
+        {
+            correctedStartBytes = new byte[endBytes.length];
+        }
+        else if (startBytes.length != endBytes.length)
         {
             throw new IllegalArgumentException(
                 "The start and the end values must not have different lengths.");
         }
-        start = new int[startBytes.length];
+        start = new int[correctedStartBytes.length];
         end = new int[endBytes.length];
-        for (int i = 0; i < startBytes.length; i++)
+        for (int i = 0; i < correctedStartBytes.length; i++)
         {
-            start[i] = startBytes[i] & 0xFF;
+            start[i] = correctedStartBytes[i] & 0xFF;
             end[i] = endBytes[i] & 0xFF;
         }
         codeLength = endBytes.length;

diff --git a/library/src/main/java/com/tom_roush/fontbox/pfb/PfbParser.java b/library/src/main/java/com/tom_roush/fontbox/pfb/PfbParser.java
@@ -140,11 +140,21 @@ private void parsePfb(final byte[] pfb) throws IOException
             size += in.read() << 8;
             size += in.read() << 16;
             size += in.read() << 24;
+            if (size < 0)
+            {
+                throw new IOException("PFB record size is negative: " + size);
+            }
             lengths[records] = size;
             if (pointer >= pfbdata.length)
             {
                 throw new EOFException("attempted to read past EOF");
             }
+            if (size > pfbdata.length - pointer)
+            {
+                throw new IOException("PFB record size (" + size +
+                    ") doesn't fit in buffer, position: " + pointer +
+                    ", total length: " + pfbdata.length);
+            }
             int got = in.read(pfbdata, pointer, size);
             if (got < 0)
             {

diff --git a/library/src/main/java/com/tom_roush/fontbox/ttf/BufferedRandomAccessFile.java b/library/src/main/java/com/tom_roush/fontbox/ttf/BufferedRandomAccessFile.java
@@ -148,24 +148,37 @@ private void invalidate() throws IOException
     @Override
     public int read(byte[] b, int off, int len) throws IOException
     {
-        int leftover = bufend - bufpos;
-        if (len <= leftover)
-        {
-            System.arraycopy(buffer, bufpos, b, off, len);
-            bufpos += len;
-            return len;
-        }
-        System.arraycopy(buffer, bufpos, b, off, leftover);
-        bufpos += leftover;
-        if (fillBuffer() > 0)
+        int curLen = len; // length of what is left to read (shrinks)
+        int curOff = off; // offset where to put read data (grows)
+        int totalRead = 0;
+
+        while (true)
         {
-            int bytesRead = read(b, off + leftover, len - leftover);
-            if (bytesRead > 0)
+            int leftover = bufend - bufpos;
+            if (curLen <= leftover)
+            {
+                System.arraycopy(buffer, bufpos, b, curOff, curLen);
+                bufpos += curLen;
+                return totalRead + curLen;
+            }
+            // curLen > leftover, we need to read more than what remains in buffer
+            System.arraycopy(buffer, bufpos, b, curOff, leftover);
+            totalRead += leftover;
+            bufpos += leftover;
+            if (fillBuffer() > 0)
+            {
+                curOff += leftover;
+                curLen -= leftover;
+            }
+            else
             {
-                leftover += bytesRead;
+                if (totalRead == 0)
+                {
+                    return -1;
+                }
+                return totalRead;
             }
         }
-        return leftover > 0 ? leftover : -1;
     }
 
     /**

diff --git a/library/src/main/java/com/tom_roush/fontbox/ttf/GlyphSubstitutionTable.java b/library/src/main/java/com/tom_roush/fontbox/ttf/GlyphSubstitutionTable.java
@@ -402,7 +402,7 @@ private Collection<LangSysTable> getLangSysTables(String scriptTag)
      *
      * @param langSysTables The {@code LangSysTable}s indicating {@code FeatureRecord}s to search
      * for
-     * @param enabledFeatures An optional whitelist of feature tags ({@code null} to allow all)
+     * @param enabledFeatures An optional list of feature tags ({@code null} to allow all)
      * @return The indicated {@code FeatureRecord}s
      */
     private List<FeatureRecord> getFeatureRecords(Collection<LangSysTable> langSysTables,
@@ -510,16 +510,16 @@ private int doLookup(LookupTable lookupTable, int gid)
 
     /**
      * Apply glyph substitutions to the supplied gid. The applicable substitutions are determined by
-     * the {@code scriptTags} which indicate the language of the gid, and by the
-     * {@code enabledFeatures} which acts as a whitelist.
+     * the {@code scriptTags} which indicate the language of the gid, and by the list of
+     * {@code enabledFeatures}.
      *
      * To ensure that a single gid isn't mapped to multiple substitutions, subsequent invocations
      * with the same gid will return the same result as the first, regardless of script or enabled
      * features.
      *
      * @param gid GID
      * @param scriptTags Script tags applicable to the gid (see {@link OpenTypeScript})
-     * @param enabledFeatures Whitelist of features to apply
+     * @param enabledFeatures list of features to apply
      */
     public int getSubstitution(int gid, String[] scriptTags, List<String> enabledFeatures)
     {

diff --git a/library/src/main/java/com/tom_roush/fontbox/ttf/OS2WindowsMetricsTable.java b/library/src/main/java/com/tom_roush/fontbox/ttf/OS2WindowsMetricsTable.java
@@ -152,7 +152,7 @@ public class OS2WindowsMetricsTable extends TTFTable
      * <p>For Restricted License embedding to take effect, it must be the only level of embedding
      * selected.
      */
-    public static final short FSTYPE_RESTRICTED = 0x0001;
+    public static final short FSTYPE_RESTRICTED = 0x0002;
 
     /**
      * Preview and Print embedding: the font may be embedded, and temporarily loaded on the

diff --git a/library/src/main/java/com/tom_roush/fontbox/ttf/OpenTypeScript.java b/library/src/main/java/com/tom_roush/fontbox/ttf/OpenTypeScript.java
@@ -18,6 +18,7 @@
 
 import android.util.Log;
 
+import java.io.BufferedInputStream;
 import java.io.IOException;
 import java.io.InputStream;
 import java.io.InputStreamReader;
@@ -222,21 +223,13 @@ public final class OpenTypeScript
         {
             if (PDFBoxResourceLoader.isReady())
             {
-                input = PDFBoxResourceLoader.getStream(path);
+                input = new BufferedInputStream(PDFBoxResourceLoader.getStream(path));
             }
             else
             {
-                // Fallback
-                input = OpenTypeScript.class.getResourceAsStream(path);
-            }
-            if (input != null)
-            {
-                parseScriptsFile(input);
-            }
-            else
-            {
-                Log.w("PdfBox-Android", "Could not find '" + path + "', mirroring char map will be empty: ");
+                input = new BufferedInputStream(OpenTypeScript.class.getResourceAsStream(path));
             }
+            parseScriptsFile(input);
         }
         catch (IOException e)
         {