Accept extraneous space before xref (fixes #341)

Jay Berkenbilt
1 parent 8a9086a6
Showing 10 changed files with 228 additions and 1 deletions
ChangeLog
libqpdf/QPDF.cc
qpdf/qpdf.testcov
qpdf/qtest/qpdf.test
qpdf/qtest/qpdf/bad37-recover.out
qpdf/qtest/qpdf/bad37.out
qpdf/qtest/qpdf/bad37.pdf
qpdf/qtest/qpdf/bad38-recover.out
qpdf/qtest/qpdf/bad38.out
qpdf/qtest/qpdf/bad38.pdf
 2019-08-19  Jay Berkenbilt  <ejb@ql.org>
  
+	* Accept (and warn for) extraneous whitespace preceding the xref
+	table. Fixes #341.
+
 	* Accept (and warn for) extraneous whitespace between the stream
 	keyword and newline. Fixes #329.
  
@@ -570,6 +570,37 @@ QPDF::read_xref(qpdf_offset_t xref_offset)
         char buf[7];
         memset(buf, 0, sizeof(buf));
 	this->m->file->seek(xref_offset, SEEK_SET);
+        // Some files miss the mark a little with startxref. We could
+        // do a better job of searching in the neighborhood for
+        // something that looks like either an xref table or stream,
+        // but the simple heuristic of skipping whitespace can help
+        // with the xref table case and is harmless with the stream
+        // case.
+        bool done = false;
+        bool skipped_space = false;
+        while (! done)
+        {
+            char ch;
+            if (1 == this->m->file->read(&ch, 1))
+            {
+                if (QUtil::is_space(ch))
+                {
+                    skipped_space = true;
+                }
+                else
+                {
+                    this->m->file->unreadCh(ch);
+                    done = true;
+                }
+            }
+            else
+            {
+                QTC::TC("qpdf", "QPDF eof skipping spaces before xref",
+                        skipped_space ? 0 : 1);
+                done = true;
+            }
+        }
+
 	this->m->file->read(buf, sizeof(buf) - 1);
         // The PDF spec says xref must be followed by a line
         // terminator, but files exist in the wild where it is
@@ -577,6 +608,13 @@ QPDF::read_xref(qpdf_offset_t xref_offset)
         if ((strncmp(buf, "xref", 4) == 0) &&
             QUtil::is_space(buf[4]))
 	{
+            if (skipped_space)
+            {
+                QTC::TC("qpdf", "QPDF xref skipped space");
+                warn(QPDFExc(qpdf_e_damaged_pdf, this->m->file->getName(),
+                             "", 0,
+                             "extraneous whitespace seen before xref"));
+            }
             QTC::TC("qpdf", "QPDF xref space",
                     ((buf[4] == '\n') ? 0 :
                      (buf[4] == '\r') ? 1 :
@@ -441,3 +441,5 @@ QPDFObjectHandle int returning INT_MIN 0
 QPDFObjectHandle int returning INT_MAX 0
 QPDFObjectHandle uint returning UINT_MAX 0
 QPDFObjectHandle uint uint returning 0 0
+QPDF xref skipped space 0
+QPDF eof skipping spaces before xref 1
@@ -2291,6 +2291,8 @@ my @badfiles = (&quot;not a PDF file&quot;, 			# 1
 		"obj/gen in wrong place",		# 34
 		"object stream of wrong type",		# 35
                 "bad dictionary key",                   # 36
+                "space before xref",                    # 37
+                "startxref to space then eof",          # 38
 		);
  
 $n_tests += @badfiles + 6;
@@ -2301,7 +2303,7 @@ $n_tests += @badfiles + 6;
 # have error conditions that used to be fatal but are now considered
 # non-fatal.
 my %badtest_overrides = ();
-for(6, 12..15, 17, 18..32, 34, 36)
+for(6, 12..15, 17, 18..32, 34, 36..37)
 {
     $badtest_overrides{$_} = 0;
 }
+WARNING: bad37.pdf: extraneous whitespace seen before xref
+/QTest is direct and has type string (6)
+/QTest is a string with value potato
+unparse: (potato)
+unparseResolved: (potato)
+test 1 done
+WARNING: bad37.pdf: extraneous whitespace seen before xref
+/QTest is direct and has type string (6)
+/QTest is a string with value potato
+unparse: (potato)
+unparseResolved: (potato)
+test 0 done
+%PDF-1.3
+1 0 obj
+<<
+  /Type /Catalog
+  /Pages 2 0 R
+>>
+endobj
+
+2 0 obj
+<<
+  /Type /Pages
+  /Kids [
+    3 0 R
+  ]
+  /Count 1
+>>
+endobj
+
+3 0 obj
+<<
+  /Type /Page
+  /Parent 2 0 R
+  /MediaBox [0 0 612 792]
+  /Contents 4 0 R
+  /Resources <<
+    /ProcSet 5 0 R
+    /Font <<
+      /F1 6 0 R
+    >>
+  >>
+>>
+endobj
+
+4 0 obj
+<<
+  /Length 44
+>>
+stream
+BT
+  /F1 24 Tf
+  72 720 Td
+  (Potato) Tj
+ET
+endstream
+endobj
+
+5 0 obj
+[
+  /PDF
+  /Text
+]
+endobj
+
+6 0 obj
+<<
+  /Type /Font
+  /Subtype /Type1
+  /Name /F1
+  /BaseFont /Helvetica
+  /Encoding /WinAnsiEncoding
+>>
+endobj
+
+xref
+0 7
+0000000000 65535 f 
+0000000009 00000 n 
+0000000063 00000 n 
+0000000135 00000 n 
+0000000307 00000 n 
+0000000403 00000 n 
+0000000438 00000 n 
+trailer <<
+  /Size 7
+  /Root 1 0 R
+  /QTest (potato)
+>>
+startxref
+555
+%%EOF
+WARNING: bad38.pdf: file is damaged
+WARNING: bad38.pdf (offset 781): xref not found
+WARNING: bad38.pdf: Attempting to reconstruct cross-reference table
+/QTest is direct and has type string (6)
+/QTest is a string with value potato
+unparse: (potato)
+unparseResolved: (potato)
+test 1 done
+bad38.pdf (offset 781): xref not found
+%PDF-1.3
+1 0 obj
+<<
+  /Type /Catalog
+  /Pages 2 0 R
+>>
+endobj
+
+2 0 obj
+<<
+  /Type /Pages
+  /Kids [
+    3 0 R
+  ]
+  /Count 1
+>>
+endobj
+
+3 0 obj
+<<
+  /Type /Page
+  /Parent 2 0 R
+  /MediaBox [0 0 612 792]
+  /Contents 4 0 R
+  /Resources <<
+    /ProcSet 5 0 R
+    /Font <<
+      /F1 6 0 R
+    >>
+  >>
+>>
+endobj
+
+4 0 obj
+<<
+  /Length 44
+>>
+stream
+BT
+  /F1 24 Tf
+  72 720 Td
+  (Potato) Tj
+ET
+endstream
+endobj
+
+5 0 obj
+[
+  /PDF
+  /Text
+]
+endobj
+
+6 0 obj
+<<
+  /Type /Font
+  /Subtype /Type1
+  /Name /F1
+  /BaseFont /Helvetica
+  /Encoding /WinAnsiEncoding
+>>
+endobj
+
+xref
+0 7
+0000000000 65535 f 
+0000000009 00000 n 
+0000000063 00000 n 
+0000000135 00000 n 
+0000000307 00000 n 
+0000000403 00000 n 
+0000000438 00000 n 
+trailer <<
+  /Size 7
+  /Root 1 0 R
+  /QTest (potato)
+>>
+startxref
+781
+%%EOF
+