Fix 5953

pdfbox/pom.xml

@@ -1098,6 +1098,19 @@
                             <sha512>f3d0934c9d0babedd76431565bdb1cc1bf6d9f0a5ace7656a9ded53fb1e701dd0d12c784a58ae87098793f57ba5b98cbfb657e047e27d5464b26634511162264</sha512>
                         </configuration>
                     </execution>
+                    <execution>
+                        <id>PDFBOX-5953</id>
+                        <phase>generate-test-resources</phase>
+                        <goals>
+                            <goal>wget</goal>
+                        </goals>
+                        <configuration>
+                            <url>https://issues.apache.org/jira/secure/attachment/13074663/test.pdf</url>
+                            <outputDirectory>${project.build.directory}/pdfs</outputDirectory>
+                            <outputFileName>PDFBOX-5953-test.pdf</outputFileName>
+                            <sha512>022755d4abeaa60ed99b3252441aaa1b65daa8d5414348f6c4ae5c8e3dfdb7d2be5ea0705117b8bfbb88f13e79d168dcb1f3b2ff9d9c7be21ad987965ccfc562</sha512>
+                        </configuration>
+                    </execution>
                 </executions>
             </plugin>
         </plugins>

pdfbox/src/main/java/org/apache/pdfbox/pdmodel/font/PDCIDFontType2.java

@@ -398,6 +398,20 @@ protected byte[] encode(int unicode, PDType0Font parent)
                 cid = 0;
             }
         }
+        else if (!parent.getCMap().getName().startsWith("Identity-") &&
+                 parent.getCMap().getName().startsWith("Uni") &&
+                 parent.getCMapUCS2() != null)
+        {
+            // PDFBOX-5953: predefined "Uni...-UCS2/UTF16" CMaps use the raw UTF-16BE
+            // Unicode value as their codespace/code by definition (see e.g. the
+            // UniGB-UTF16-H resource, whose begincidchar entries map codes that are
+            // themselves Unicode code points, e.g. <00a4> to a CID). Writing the
+            // substituted font's own glyph index here instead (as done below for the
+            // general case) produces a code that decodes to a wrong, essentially
+            // arbitrary character once the CID/Unicode round trip in
+            // PDCIDFontType2#codeToGID or PDType0Font#toUnicode is applied to it.
+            cid = unicode;
+        }
         else
         {
             // a non-embedded font always has a cmap (otherwise it we wouldn't load it)

pdfbox/src/test/java/org/apache/pdfbox/pdmodel/font/PDCIDFontType2SubstituteTest.java

@@ -0,0 +1,84 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements.  See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You under the Apache License, Version 2.0
+ * (the "License"); you may not use this file except in compliance with
+ * the License.  You may obtain a copy of the License at
+ *
+ *      http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+package org.apache.pdfbox.pdmodel.font;
+
+import static org.junit.jupiter.api.Assertions.assertEquals;
+import static org.junit.jupiter.api.Assumptions.assumeTrue;
+
+import java.io.File;
+import java.io.IOException;
+
+import org.apache.fontbox.ttf.CmapLookup;
+import org.apache.pdfbox.Loader;
+import org.apache.pdfbox.cos.COSName;
+import org.apache.pdfbox.pdmodel.PDDocument;
+import org.apache.pdfbox.pdmodel.interactive.form.PDAcroForm;
+
+import org.junit.jupiter.api.Test;
+
+/**
+ * A non-embedded CIDFontType2 whose declared /Encoding is a predefined "Uni...-UCS2/UTF16" CMap
+ * must write a code that is valid under that encoding - the raw Unicode value - when generating
+ * new appearance content, not the substituted font's own glyph index. Before the fix, the
+ * substitute font's glyph index was written regardless of the declared encoding; since that
+ * index is unrelated to the Unicode-based codespace of a "Uni...-UTF16-H" CMap, a conforming
+ * reader (including PDFBox itself) decoded it back into a wrong, essentially arbitrary
+ * character. PDFBOX-5953.
+ *
+ * <p>The fixture is the AcroForm/DR "SimSun" font (Type0/CIDFontType2, /Encoding
+ * UniGB-UTF16-H, not embedded) of a real-world Chinese bank receipt form attached to the JIRA
+ * issue, whose filled-in field values (account numbers, amounts, company names) render as
+ * garbled or blank text because of this bug.
+ */
+class PDCIDFontType2SubstituteTest
+{
+    @Test
+    void testEncodeNonEmbeddedPredefinedUnicodeEncoding() throws IOException
+    {
+        File file = new File("target/pdfs", "PDFBOX-5953-test.pdf");
+        try (PDDocument doc = Loader.loadPDF(file))
+        {
+            PDAcroForm acroForm = doc.getDocumentCatalog().getAcroForm();
+            PDType0Font font = (PDType0Font) acroForm.getDefaultResources()
+                    .getFont(COSName.getPDFName("SimSun"));
+            PDCIDFontType2 cidFont = (PDCIDFontType2) font.getDescendantFont();
+            assumeTrue(!cidFont.isEmbedded(),
+                    "font is embedded in this fixture, can't test substitution");
+            assumeTrue(font.getCMap().getName().startsWith("Uni"),
+                    "fixture font's encoding changed, can't test");
+
+            CmapLookup substituteCmap = cidFont.getTrueTypeFont().getUnicodeCmapLookup(false);
+            // '百' (as in 百色, the city in the receipts): a common GB1 Hanzi, guaranteed to be
+            // resolvable through the Adobe-GB1 <-> Unicode predefined CMaps
+            int unicode = 0x767E;
+            int expectedGid = substituteCmap.getGlyphId(unicode);
+            assumeTrue(expectedGid != 0,
+                    "no CJK glyph for U+767E in the substituted font, can't test");
+
+            byte[] bytes = cidFont.encode(unicode, font);
+            assertEquals(2, bytes.length);
+            int code = ((bytes[0] & 0xff) << 8) | (bytes[1] & 0xff);
+
+            // before the fix, 'code' was the substitute font's own glyph index; decoding it
+            // again (as any reader must, since the font's /Encoding is UniGB-UTF16-H, not
+            // Identity-H) produced a wrong, essentially arbitrary glyph instead of round
+            // tripping back to the glyph for U+767E
+            assertEquals(expectedGid, cidFont.codeToGID(code, font),
+                    "encoded code must decode back to the substitute font's glyph for U+767E");
+        }
+    }
+}