Fix linearization for files with nonstandard ID length

Jay Berkenbilt
1 parent 9732de7e
Showing 7 changed files with 72 additions and 10 deletions
include/qpdf/QPDFWriter.hh
libqpdf/QPDFWriter.cc
qpdf/qtest/qpdf.test
qpdf/qtest/qpdf/long-id-linearized-check.out
qpdf/qtest/qpdf/long-id-linearized.pdf
qpdf/qtest/qpdf/short-id-linearized-check.out
qpdf/qtest/qpdf/short-id-linearized.pdf
@@ -414,6 +414,7 @@ class QPDFWriter
     void initializeSpecialStreams();
     void preserveObjectStreams();
     void generateObjectStreams();
+    std::string getOriginalID1();
     void generateID();
     void interpretR3EncryptionParameters(
 	std::set<int>& bits_to_clear,
@@ -1246,8 +1246,28 @@ QPDFWriter::writeTrailer(trailer_e which, int size, bool xref_stream,
     writeString(" /ID [");
     if (linearization_pass == 1)
     {
-        writeString("<00000000000000000000000000000000>"
-                    "<00000000000000000000000000000000>");
+	std::string original_id1 = getOriginalID1();
+        if (original_id1.empty())
+        {
+            writeString("<00000000000000000000000000000000>");
+        }
+        else
+        {
+            // Write a string of zeroes equal in length to the
+            // representation of the original ID. While writing the
+            // original ID would have the same number of bytes, it
+            // would cause a change to the deterministic ID generated
+            // by older versions of the software that hard-coded the
+            // length of the ID to 16 bytes.
+            writeString("<");
+            size_t len = QPDF_String(original_id1).unparse(true).length() - 2;
+            for (size_t i = 0; i < len; ++i)
+            {
+                writeString("0");
+            }
+            writeString(">");
+        }
+        writeString("<00000000000000000000000000000000>");
     }
     else
     {
@@ -1952,6 +1972,20 @@ QPDFWriter::writeObject(QPDFObjectHandle object, int object_stream_index)
     }
 }
  
+std::string
+QPDFWriter::getOriginalID1()
+{
+    QPDFObjectHandle trailer = this->m->pdf.getTrailer();
+    if (trailer.hasKey("/ID"))
+    {
+        return trailer.getKey("/ID").getArrayItem(0).getStringValue();
+    }
+    else
+    {
+        return "";
+    }
+}
+
 void
 QPDFWriter::generateID()
 {
@@ -2042,12 +2076,9 @@ QPDFWriter::generateID()
     // generated ID for both.
  
     this->m->id2 = result;
-    if (trailer.hasKey("/ID"))
-    {
-	// Note: keep /ID from old file even if --static-id was given.
-	this->m->id1 = trailer.getKey("/ID").getArrayItem(0).getStringValue();
-    }
-    else
+    // Note: keep /ID from old file even if --static-id was given.
+    this->m->id1 = getOriginalID1();
+    if (this->m->id1.empty())
     {
 	this->m->id1 = this->m->id2;
     }
@@ -240,7 +240,7 @@ foreach my $d (@bug_tests)
 show_ntests();
 # ----------
 $td->notify("--- Miscellaneous Tests ---");
-$n_tests += 87;
+$n_tests += 93;
  
 $td->runtest("qpdf version",
 	     {$td->COMMAND => "qpdf --version"},
@@ -303,7 +303,8 @@ $td-&gt;runtest(&quot;check output&quot;,
 	     {$td->FILE => "unreferenced-indirect-scalar.out"});
  
 # Encrypt files whose /ID strings are other than 32 bytes long (bug
-# 2991412).
+# 2991412). Also linearize these files, which was reported in a
+# separate bug by email.
 foreach my $file (qw(short-id long-id))
 {
     $td->runtest("encrypt $file.pdf",
@@ -318,6 +319,23 @@ foreach my $file (qw(short-id long-id))
 		 {$td->FILE => "$file-check.out",
 		  $td->EXIT_STATUS => 0},
 		 $td->NORMALIZE_NEWLINES);
+
+    $td->runtest("linearize $file.pdf",
+		 {$td->COMMAND =>
+		      "qpdf --deterministic-id --linearize $file.pdf a.pdf"},
+		 {$td->STRING => "",
+		  $td->EXIT_STATUS => 0},
+		 $td->NORMALIZE_NEWLINES);
+
+    $td->runtest("check output",
+                 {$td->FILE => "a.pdf"},
+		 {$td->FILE => "$file-linearized.pdf"});
+
+    $td->runtest("check $file.pdf",
+		 {$td->COMMAND => "qpdf --check a.pdf"},
+		 {$td->FILE => "$file-linearized-check.out",
+		  $td->EXIT_STATUS => 0},
+		 $td->NORMALIZE_NEWLINES);
 }
  
 # Handle file with invalid xref table and object 0 as a regular object
+checking a.pdf
+PDF Version: 1.3
+File is not encrypted
+File is linearized
+No syntax or stream encoding errors found; the file may still contain
+errors that qpdf cannot detect
+checking a.pdf
+PDF Version: 1.3
+File is not encrypted
+File is linearized
+No syntax or stream encoding errors found; the file may still contain
+errors that qpdf cannot detect