accept stream keyword with CR onlyrelease-qpdf-2.2.3

git-svn-id: svn+q:///qpdf/trunk@1052 71b93d88-0707-0410-a8cf-f5a4172ac649

8 files changed, 272 insertions, 14 deletions

diff --git a/ChangeLog b/ChangeLog
index 01e43879..a81124bf 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -2,6 +2,11 @@
 
 	* 2.2.3: release
 
+	* libqpdf/QPDF.cc (readObjectInternal): Accept the case of the
+	stream keyword being followed by carriage return by itself.  While
+	this is not permitted by the specification, there are PDF files
+	that do this, and other readers can read them.
+
 	* libqpdf/Pl_QPDFTokenizer.cc (processChar): When an inline image
 	is detected, suspend normalization only up to the end of the
 	inline image rather than for the remainder of the content stream.
diff --git a/libqpdf/QPDF.cc b/libqpdf/QPDF.cc
index c8146eff..f6157287 100644
--- a/libqpdf/QPDF.cc
+++ b/libqpdf/QPDF.cc
@@ -1331,24 +1331,66 @@ QPDF::readObjectInternal(PointerHolder<InputSource> input,
 	    if (readToken(input) ==
 		QPDFTokenizer::Token(QPDFTokenizer::tt_word, "stream"))
 	    {
-		// Kill to next actual newline.  Do not use readLine()
-		// here -- streams are a special case.  The next
-		// single newline character marks the end of the
-		// stream token.  It is incorrect to strip subsequent
-		// carriage returns or newlines as they may be part of
-		// the stream.
+		// The PDF specification states that the word "stream"
+		// should be followed by either a carriage return and
+		// a newline or by a newline alone.  It specifically
+		// disallowed following it by a carriage return alone
+		// since, in that case, there would be no way to tell
+		// whether the NL in a CR NL sequence was part of the
+		// stream data.  However, some readers, including
+		// Adobe reader, accept a carriage return by itself
+		// when followed by a non-newline character, so that's
+		// what we do here.
 		{
 		    char ch;
-		    do
+		    if (input->read(&ch, 1) == 0)
 		    {
-			if (input->read(&ch, 1) == 0)
+			// A premature EOF here will result in some
+			// other problem that will get reported at
+			// another time.
+		    }
+		    else if (ch == '\n')
+		    {
+			// ready to read stream data
+			QTC::TC("qpdf", "QPDF stream with NL only");
+		    }
+		    else if (ch == '\r')
+		    {
+			// Read another character
+			if (input->read(&ch, 1) != 0)
 			{
-			    // A premature EOF here will result in
-			    // some other problem that will get
-			    // reported at another time.
-			    ch = '\n';
+			    if (ch == '\n')
+			    {
+				// Ready to read stream data
+				QTC::TC("qpdf", "QPDF stream with CRNL");
+			    }
+			    else
+			    {
+				// Treat the \r by itself as the
+				// whitespace after endstream and
+				// start reading stream data in spite
+				// of not having seen a newline.
+				QTC::TC("qpdf", "QPDF stream with CR only");
+				input->unreadCh(ch);
+				warn(QPDFExc(
+					 qpdf_e_damaged_pdf,
+					 input->getName(),
+					 this->last_object_description,
+					 input->tell(),
+					 "stream keyword followed"
+					 " by carriage return only"));
+			    }
 			}
-		    } while (ch != '\n');
+		    }
+		    else
+		    {
+			QTC::TC("qpdf", "QPDF stream without newline");
+			warn(QPDFExc(qpdf_e_damaged_pdf, input->getName(),
+				     this->last_object_description,
+				     input->tell(),
+				     "stream keyword not followed"
+				     " by proper line terminator"));
+		    }
 		}
 
 		// Must get offset before accessing any additional
diff --git a/manual/qpdf-manual.xml b/manual/qpdf-manual.xml
index 9bdb95b3..02b69a1d 100644
--- a/manual/qpdf-manual.xml
+++ b/manual/qpdf-manual.xml
@@ -2080,6 +2080,12 @@ print "\n";
      <itemizedlist>
       <listitem>
        <para>
+        Handle some damaged streams with incorrect characters
+        following the stream keyword.
+       </para>
+      </listitem>
+      <listitem>
+       <para>
         Improve handling of inline images when normalizing content
         streams.
        </para>
diff --git a/qpdf/qpdf.testcov b/qpdf/qpdf.testcov
index 51a6f0e8..4e7e292a 100644
--- a/qpdf/qpdf.testcov
+++ b/qpdf/qpdf.testcov
@@ -188,3 +188,7 @@ QPDF_Stream getStreamData 0
 QPDF_Stream expand filter abbreviation 0
 qpdf-c called qpdf_read_memory 0
 Pl_QPDFTokenizer found EI 0
+QPDF stream without newline 0
+QPDF stream with CR only 0
+QPDF stream with CRNL 0
+QPDF stream with NL only 0
diff --git a/qpdf/qtest/qpdf.test b/qpdf/qtest/qpdf.test
index 62eed000..75a92aa4 100644
--- a/qpdf/qtest/qpdf.test
+++ b/qpdf/qtest/qpdf.test
@@ -111,7 +111,7 @@ $td->runtest("new stream",
 show_ntests();
 # ----------
 $td->notify("--- Miscellaneous Tests ---");
-$n_tests += 29;
+$n_tests += 31;
 
 $td->runtest("qpdf version",
 	     {$td->COMMAND => "qpdf --version"},
@@ -265,6 +265,17 @@ $td->runtest("error/output redirection to strings",
 	      $td->EXIT_STATUS => 0},
 	     $td->NORMALIZE_NEWLINES);
 
+$td->runtest("odd terminators for stream keyword",
+	     {$td->COMMAND =>
+		  "qpdf --qdf --static-id" .
+		  " stream-line-enders.pdf a.qdf"},
+	     {$td->FILE => "stream-line-enders.out",
+	      $td->EXIT_STATUS => 3},
+	     $td->NORMALIZE_NEWLINES);
+$td->runtest("check output",
+	     {$td->FILE => "a.qdf"},
+	     {$td->FILE => "stream-line-enders.qdf"});
+
 show_ntests();
 # ----------
 $td->notify("--- Error Condition Tests ---");
diff --git a/qpdf/qtest/qpdf/stream-line-enders.out b/qpdf/qtest/qpdf/stream-line-enders.out
new file mode 100644
index 00000000..b7a7513c
--- /dev/null
+++ b/qpdf/qtest/qpdf/stream-line-enders.out
@@ -0,0 +1,3 @@
+WARNING: stream-line-enders.pdf (object 5 0, file position 378): stream keyword followed by carriage return only
+WARNING: stream-line-enders.pdf (object 6 0, file position 437): stream keyword not followed by proper line terminator
+qpdf: operation succeeded with warnings; resulting file may have some problems
diff --git a/qpdf/qtest/qpdf/stream-line-enders.pdf b/qpdf/qtest/qpdf/stream-line-enders.pdf
new file mode 100644
index 00000000..e6238889
--- /dev/null
+++ b/qpdf/qtest/qpdf/stream-line-enders.pdf
@@ -0,0 +1,50 @@
+%PDF-1.3
+%¿÷¢þ
+1 0 obj
+<< /Pages 2 0 R /Type /Catalog >>
+endobj
+2 0 obj
+<< /Count 1 /Kids [ 3 0 R ] /Type /Pages >>
+endobj
+3 0 obj
+<< /Contents [ 4 0 R 5 0 R 6 0 R ] /MediaBox [ 0 0 612 792 ] /Parent 2 0 R /Resources << /Font << /F1 7 0 R >> /ProcSet 8 0 R >> /Type /Page >>
+endobj
+4 0 obj
+<< /Length 14 >>
+stream

+BT
+ /F1 24 Tf
+endstream
+endobj
+5 0 obj
+<< /Length 10 >>
+stream
72 720 Td
+endstream
+endobj
+6 0 obj
+<< /Length 15 >>
+stream (Potato) Tj
+ET
+endstream
+endobj
+7 0 obj
+<< /BaseFont /Helvetica /Encoding /WinAnsiEncoding /Name /F1 /Subtype /Type1 /Type /Font >>
+endobj
+8 0 obj
+[ /PDF /Text ]
+endobj
+xref
+0 9
+0000000000 65535 f 
+0000000015 00000 n 
+0000000064 00000 n 
+0000000123 00000 n 
+0000000282 00000 n 
+0000000346 00000 n 
+0000000405 00000 n 
+0000000469 00000 n 
+0000000576 00000 n 
+trailer << /Root 1 0 R /Size 9 /ID [<08aa98c73f8a7262d77c8328772c3989><7b1f32865e2165debe277f27ee790092>] >>
+startxref
+606
+%%EOF
diff --git a/qpdf/qtest/qpdf/stream-line-enders.qdf b/qpdf/qtest/qpdf/stream-line-enders.qdf
new file mode 100644
index 00000000..33536028
--- /dev/null
+++ b/qpdf/qtest/qpdf/stream-line-enders.qdf
@@ -0,0 +1,137 @@
+%PDF-1.3
+%¿÷¢þ
+%QDF-1.0
+
+%% Original object ID: 1 0
+1 0 obj
+<<
+  /Pages 2 0 R
+  /Type /Catalog
+>>
+endobj
+
+%% Original object ID: 2 0
+2 0 obj
+<<
+  /Count 1
+  /Kids [
+    3 0 R
+  ]
+  /Type /Pages
+>>
+endobj
+
+%% Page 1
+%% Original object ID: 3 0
+3 0 obj
+<<
+  /Contents [
+    4 0 R
+    6 0 R
+    8 0 R
+  ]
+  /MediaBox [
+    0
+    0
+    612
+    792
+  ]
+  /Parent 2 0 R
+  /Resources <<
+    /Font <<
+      /F1 10 0 R
+    >>
+    /ProcSet 11 0 R
+  >>
+  /Type /Page
+>>
+endobj
+
+%% Contents for page 1
+%% Original object ID: 4 0
+4 0 obj
+<<
+  /Length 5 0 R
+>>
+stream
+BT
+ /F1 24 Tf
+endstream
+endobj
+
+5 0 obj
+14
+endobj
+
+%% Contents for page 1
+%% Original object ID: 5 0
+6 0 obj
+<<
+  /Length 7 0 R
+>>
+stream
+72 720 Td
+endstream
+endobj
+
+7 0 obj
+10
+endobj
+
+%% Contents for page 1
+%% Original object ID: 6 0
+8 0 obj
+<<
+  /Length 9 0 R
+>>
+stream
+(Potato) Tj
+ET
+endstream
+endobj
+
+9 0 obj
+15
+endobj
+
+%% Original object ID: 7 0
+10 0 obj
+<<
+  /BaseFont /Helvetica
+  /Encoding /WinAnsiEncoding
+  /Name /F1
+  /Subtype /Type1
+  /Type /Font
+>>
+endobj
+
+%% Original object ID: 8 0
+11 0 obj
+[
+  /PDF
+  /Text
+]
+endobj
+
+xref
+0 12
+0000000000 65535 f 
+0000000052 00000 n 
+0000000133 00000 n 
+0000000242 00000 n 
+0000000516 00000 n 
+0000000585 00000 n 
+0000000654 00000 n 
+0000000719 00000 n 
+0000000788 00000 n 
+0000000858 00000 n 
+0000000904 00000 n 
+0000001050 00000 n 
+trailer <<
+  /Root 1 0 R
+  /Size 12
+  /ID [<08aa98c73f8a7262d77c8328772c3989><31415926535897932384626433832795>]
+>>
+startxref
+1086
+%%EOF