Properly handle pages with no contents (fixes #194)

Remove calls to assertPageObject(). All cases in the library that
called assertPageObject() work fine if you don't call
assertPageObject() because nothing assumes anything that was being
checked by that call. Removing the calls enables more files to be
successfully processed.

ChangeLog

@@ -1,3 +1,7 @@
+2018-03-06  Jay Berkenbilt  <ejb@ql.org>
+
+	* Properly handle pages with no contents. Fixes #194.
+
 2018-03-05  Jay Berkenbilt  <ejb@ql.org>
 
 	* Improve handling of loops while following cross reference

include/qpdf/QPDFObjectHandle.hh

@@ -830,6 +830,11 @@ class QPDFObjectHandle
     QPDF_DLL
     void assertNumber();
 
+    // The isPageObject method checks the /Type key of the object.
+    // This is not completely reliable as there are some otherwise
+    // valid files whose /Type is wrong for page objects. qpdf is
+    // slightly more accepting but may still return false here when
+    // treating the object as a page would work. Use this sparingly.
     QPDF_DLL
     bool isPageObject();
     QPDF_DLL

libqpdf/QPDFObjectHandle.cc

@@ -932,8 +932,6 @@ QPDFObjectHandle::getGeneration() const
 std::map<std::string, QPDFObjectHandle>
 QPDFObjectHandle::getPageImages()
 {
-    assertPageObject();
-
     // Note: this code doesn't handle inherited resources.  If this
     // page dictionary doesn't have a /Resources key or has one whose
     // value is null or an empty dictionary, you are supposed to walk
@@ -1081,7 +1079,6 @@ QPDFObjectHandle::addPageContents(QPDFObjectHandle new_contents, bool first)
 void
 QPDFObjectHandle::rotatePage(int angle, bool relative)
 {
-    assertPageObject();
     if ((angle % 90) != 0)
     {
         throw std::runtime_error(
@@ -1137,7 +1134,6 @@ QPDFObjectHandle::rotatePage(int angle, bool relative)
 void
 QPDFObjectHandle::coalesceContentStreams()
 {
-    assertPageObject();
     QPDFObjectHandle contents = this->getKey("/Contents");
     if (contents.isStream())
     {
@@ -1218,7 +1214,6 @@ QPDFObjectHandle::parse(std::string const& object_str,
 void
 QPDFObjectHandle::pipePageContents(Pipeline* p)
 {
-    assertPageObject();
     std::string description = "page object " +
         QUtil::int_to_string(this->m->objid) + " " +
         QUtil::int_to_string(this->m->generation);
@@ -1256,7 +1251,6 @@ QPDFObjectHandle::pipeContentStreams(
 void
 QPDFObjectHandle::parsePageContents(ParserCallbacks* callbacks)
 {
-    assertPageObject();
     std::string description = "page object " +
         QUtil::int_to_string(this->m->objid) + " " +
         QUtil::int_to_string(this->m->generation);
@@ -1267,7 +1261,6 @@ QPDFObjectHandle::parsePageContents(ParserCallbacks* callbacks)
 void
 QPDFObjectHandle::filterPageContents(TokenFilter* filter, Pipeline* next)
 {
-    assertPageObject();
     std::string description = "token filter for page object " +
         QUtil::int_to_string(this->m->objid) + " " +
         QUtil::int_to_string(this->m->generation);
@@ -2222,8 +2215,29 @@ QPDFObjectHandle::assertNumber()
 bool
 QPDFObjectHandle::isPageObject()
 {
-    // Some PDF files have /Type broken on pages.
+    // See comments in QPDFObjectHandle.hh.
-    return (this->isDictionary() && this->hasKey("/Contents"));
+    if (! this->isDictionary())
+    {
+        return false;
+    }
+    if (this->hasKey("/Type"))
+    {
+        QPDFObjectHandle type = this->getKey("/Type");
+        if (type.isName() && (type.getName() == "/Page"))
+        {
+            return true;
+        }
+        // Files have been seen in the wild that have /Type (Page)
+        if (type.isString() && (type.getStringValue() == "Page"))
+        {
+            return true;
+        }
+    }
+    if (this->hasKey("/Contents"))
+    {
+        return true;
+    }
+    return false;
 }
 
 bool

libqpdf/QPDF_pages.cc

@@ -191,7 +191,6 @@ QPDF::insertPage(QPDFObjectHandle newpage, int pos)
     // pos = npages adds to the end.
 
     flattenPagesTree();
-    newpage.assertPageObject();
 
     if (! newpage.isIndirect())
     {
@@ -288,7 +287,6 @@ QPDF::addPage(QPDFObjectHandle newpage, bool first)
 int
 QPDF::findPage(QPDFObjectHandle& page)
 {
-    page.assertPageObject();
     return findPage(page.getObjGen());
 }
 

qpdf/qtest/qpdf.test

@@ -919,6 +919,34 @@ $td->runtest("check output",
              {$td->FILE => "a.pdf"},
              {$td->FILE => "coalesce-out.pdf"});
 
+show_ntests();
+# ----------
+$td->notify("--- Page with no contents ---");
+$n_tests += 7;
+
+$td->runtest("check no contents",
+             {$td->COMMAND => "qpdf --check no-contents.pdf"},
+             {$td->FILE => "no-contents-check.out", $td->EXIT_STATUS => 0},
+             $td->NORMALIZE_NEWLINES);
+
+foreach my $arg ('--qdf', '--coalesce-contents', '')
+{
+    $td->runtest("convert no contents ($arg)",
+                 {$td->COMMAND =>
+                      "qpdf $arg --static-id no-contents.pdf a.pdf"},
+                 {$td->STRING => "", $td->EXIT_STATUS => 0});
+
+    my $suf = $arg;
+    $suf =~ s/--//;
+    if ($suf eq '')
+    {
+        $suf = "none";
+    }
+    $td->runtest("check output",
+                 {$td->FILE => "a.pdf"},
+                 {$td->FILE => "no-contents-$suf.pdf"});
+}
+
 show_ntests();
 # ----------
 $td->notify("--- Token filters ---");

qpdf/qtest/qpdf/no-contents-check.out

@@ -0,0 +1,6 @@
+checking no-contents.pdf
+PDF Version: 1.3
+File is not encrypted
+File is not linearized
+No syntax or stream encoding errors found; the file may still contain
+errors that qpdf cannot detect

qpdf/qtest/qpdf/no-contents-coalesce-contents.pdf

@@ -0,0 +1,27 @@
+%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 /MediaBox [ 0 0 720 720 ] /Parent 2 0 R /Resources << >> /Type /Page >>
+endobj
+4 0 obj
+<< /Length 0 /Filter /FlateDecode >>
+stream
+endstream
+endobj
+xref
+0 5
+0000000000 65535 f 
+0000000015 00000 n 
+0000000064 00000 n 
+0000000123 00000 n 
+0000000229 00000 n 
+trailer << /Root 1 0 R /Size 5 /ID [<52bba3c78160d0c6e851b59110e5d076><31415926535897932384626433832795>] >>
+startxref
+298
+%%EOF

qpdf/qtest/qpdf/no-contents-none.pdf

@@ -0,0 +1,21 @@
+%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
+<< /MediaBox [ 0 0 720 720 ] /Parent 2 0 R /Resources << >> /Type /Page >>
+endobj
+xref
+0 4
+0000000000 65535 f 
+0000000015 00000 n 
+0000000064 00000 n 
+0000000123 00000 n 
+trailer << /Root 1 0 R /Size 4 /ID [<52bba3c78160d0c6e851b59110e5d076><31415926535897932384626433832795>] >>
+startxref
+213
+%%EOF

qpdf/qtest/qpdf/no-contents-qdf.pdf

@@ -0,0 +1,54 @@
+%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
+<<
+  /MediaBox [
+    0
+    0
+    720
+    720
+  ]
+  /Parent 2 0 R
+  /Resources <<
+  >>
+  /Type /Page
+>>
+endobj
+
+xref
+0 4
+0000000000 65535 f 
+0000000052 00000 n 
+0000000133 00000 n 
+0000000242 00000 n 
+trailer <<
+  /Root 1 0 R
+  /Size 4
+  /ID [<52bba3c78160d0c6e851b59110e5d076><31415926535897932384626433832795>]
+>>
+startxref
+361
+%%EOF

qpdf/qtest/qpdf/no-contents.pdf

@@ -0,0 +1,21 @@
+%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
+<< /MediaBox [ 0 0 720 720 ] /Parent 2 0 R /Resources << >> /Type /Page >>
+endobj
+xref
+0 4
+0000000000 65535 f 
+0000000015 00000 n 
+0000000064 00000 n 
+0000000123 00000 n 
+trailer << /Root 1 0 R /Size 4 /ID [<52bba3c78160d0c6e851b59110e5d076><52bba3c78160d0c6e851b59110e5d076>] >>
+startxref
+213
+%%EOF