Attempt to repair /Type key in pages nodes (fixes #349)

ChangeLog

@@ -1,3 +1,10 @@
+2019-08-18  Jay Berkenbilt  <ejb@ql.org>
+
+	* When traversing the pages tree, if an invalid /Type key is
+	encountered, fix it. This is not done for all operations, but it
+	will be done for any case in which getAllPages is called. This
+	includes all page-based CLI operations. (Hopefully) Fixes #349.
+
 2019-08-17  Jay Berkenbilt  <ejb@ql.org>
 
 	* Change internal implementation of QPDF arrays to use sparse

libqpdf/QPDF_pages.cc

@@ -56,12 +56,12 @@ QPDF::getAllPages()
 }
 
 void
-QPDF::getAllPagesInternal(QPDFObjectHandle cur_pages,
+QPDF::getAllPagesInternal(QPDFObjectHandle cur_node,
                           std::vector<QPDFObjectHandle>& result,
                           std::set<QPDFObjGen>& visited,
                           std::set<QPDFObjGen>& seen)
 {
-    QPDFObjGen this_og = cur_pages.getObjGen();
+    QPDFObjGen this_og = cur_node.getObjGen();
     if (visited.count(this_og) > 0)
     {
         throw QPDFExc(
@@ -70,23 +70,11 @@ QPDF::getAllPagesInternal(QPDFObjectHandle cur_pages,
             "Loop detected in /Pages structure (getAllPages)");
     }
     visited.insert(this_og);
-    std::string type;
+    std::string wanted_type;
-    QPDFObjectHandle type_key = cur_pages.getKey("/Type");
+    if (cur_node.hasKey("/Kids"))
-    if (type_key.isName())
     {
-        type = type_key.getName();
+        wanted_type = "/Pages";
-    }
+	QPDFObjectHandle kids = cur_node.getKey("/Kids");
-    else if (cur_pages.hasKey("/Kids"))
-    {
-        type = "/Pages";
-    }
-    else
-    {
-        type = "/Page";
-    }
-    if (type == "/Pages")
-    {
-	QPDFObjectHandle kids = cur_pages.getKey("/Kids");
 	int n = kids.getArrayNItems();
 	for (int i = 0; i < n; ++i)
 	{
@@ -108,17 +96,22 @@ QPDF::getAllPagesInternal(QPDFObjectHandle cur_pages,
 	    getAllPagesInternal(kid, result, visited, seen);
 	}
     }
-    else if (type == "/Page")
-    {
-        seen.insert(this_og);
-	result.push_back(cur_pages);
-    }
     else
     {
-	throw QPDFExc(qpdf_e_damaged_pdf, this->m->file->getName(),
+        wanted_type = "/Page";
-		      this->m->last_object_description,
+        seen.insert(this_og);
-		      this->m->file->getLastOffset(),
+	result.push_back(cur_node);
-		      "invalid Type " + type + " in page tree");
+    }
+
+    QPDFObjectHandle type_key = cur_node.getKey("/Type");
+    if (! (type_key.isName() && (type_key.getName() == wanted_type)))
+    {
+        warn(QPDFExc(qpdf_e_damaged_pdf, this->m->file->getName(),
+                     "page tree node",
+                     this->m->file->getLastOffset(),
+                     "/Type key should be " + wanted_type +
+                     " but is not; overriding"));
+        cur_node.replaceKey("/Type", QPDFObjectHandle::newName(wanted_type));
     }
     visited.erase(this_og);
 }

manual/qpdf-manual.xml

@@ -4488,6 +4488,13 @@ print "\n";
           drastically less memory for certain types of files.
          </para>
         </listitem>
+        <listitem>
+         <para>
+          When traversing the pages tree, if nodes are encountered
+          with invalid types, the types are fixed, and a warning is
+          issued.
+         </para>
+        </listitem>
         <listitem>
          <para>
           A new helper method

qpdf/qtest/qpdf.test

@@ -1339,16 +1339,23 @@ $td->runtest("sanity check array size",
 show_ntests();
 # ----------
 $td->notify("--- Page errors ---");
-$n_tests += 3;
+$n_tests += 5;
 
 $td->runtest("handle page no with contents",
              {$td->COMMAND => "qpdf --show-pages page-no-content.pdf"},
              {$td->FILE => "page-no-content.out", $td->EXIT_STATUS => 0},
              $td->NORMALIZE_NEWLINES);
-$td->runtest("no type key for page nodes",
+$td->runtest("check no type key for page nodes",
              {$td->COMMAND => "qpdf --check no-pages-types.pdf"},
-             {$td->FILE => "no-pages-types.out", $td->EXIT_STATUS => 0},
+             {$td->FILE => "no-pages-types.out", $td->EXIT_STATUS => 3},
              $td->NORMALIZE_NEWLINES);
+$td->runtest("no type key for page nodes",
+             {$td->COMMAND => "qpdf --static-id --split-pages no-pages-types.pdf a-split-out.pdf"},
+             {$td->FILE => "no-pages-types-fix.out", $td->EXIT_STATUS => 3},
+             $td->NORMALIZE_NEWLINES);
+$td->runtest("check output",
+             {$td->FILE => "a-split-out-1.pdf"},
+             {$td->FILE => "no-pages-types-fixed.pdf"});
 $td->runtest("detect loops in pages structure",
              {$td->COMMAND => "qpdf --check pages-loop.pdf"},
              {$td->FILE => "pages-loop.out", $td->EXIT_STATUS => 2},

qpdf/qtest/qpdf/no-pages-types-fix.out

@@ -0,0 +1,3 @@
+WARNING: no-pages-types.pdf (page tree node, offset 307): /Type key should be /Page but is not; overriding
+WARNING: no-pages-types.pdf (page tree node, offset 307): /Type key should be /Pages but is not; overriding
+qpdf: operation succeeded with warnings; resulting file may have some problems

qpdf/qtest/qpdf/no-pages-types.out

@@ -2,5 +2,5 @@ checking no-pages-types.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
+WARNING: no-pages-types.pdf (page tree node, offset 307): /Type key should be /Page but is not; overriding
-errors that qpdf cannot detect
+WARNING: no-pages-types.pdf (page tree node, offset 307): /Type key should be /Pages but is not; overriding