Add C++ iterator API around array and dictionary objects

5 files changed, 427 insertions, 14 deletions

diff --git a/ChangeLog b/ChangeLog
index 4b284beb..90177d36 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,5 +1,19 @@
 2021-01-29  Jay Berkenbilt  <ejb@ql.org>
 
+	* Add wrappers QPDFDictItems and QPDFArrayItems around
+	QPDFObjectHandle that provide a C++ iterator API, including C++11
+	range-for iteration, over arrays and dictionaries. With this, you
+	can do
+
+	for (auto i: QPDFDictItems(oh))
+	{
+	    // i.first is a string, i.second is a QPDFObjectHandle
+	}
+	for (auto i: QPDFArrayItems(oh))
+	{
+	    // i is a QPDFObjectHandle
+	}
+
 	* QPDFObjectHandle::is* methods to check type now return false on
 	uninitialized objects rather than crashing or throwing a logic
 	error.
diff --git a/include/qpdf/QPDFObjectHandle.hh b/include/qpdf/QPDFObjectHandle.hh
index b28568ed..554c7131 100644
--- a/include/qpdf/QPDFObjectHandle.hh
+++ b/include/qpdf/QPDFObjectHandle.hh
@@ -631,7 +631,8 @@ class QPDFObjectHandle
     QPDF_DLL
     std::string getInlineImageValue();
 
-    // Methods for array objects; see also name and array objects
+    // Methods for array objects; see also name and array objects. See
+    // also QPDFArrayItems later in this file.
     QPDF_DLL
     int getArrayNItems();
     QPDF_DLL
@@ -655,7 +656,8 @@ class QPDFObjectHandle
     QPDF_DLL
     Matrix getArrayAsMatrix();
 
-    // Methods for dictionary objects
+    // Methods for dictionary objects. See also QPDFDictItems later in
+    // this file.
     QPDF_DLL
     bool hasKey(std::string const&);
     QPDF_DLL
@@ -1224,4 +1226,179 @@ class QPDFObjectHandle
     bool reserved;
 };
 
+class QPDFDictItems
+{
+    // This class allows C++-style iteration, including range-for
+    // iteration, around dictionaries. You can write
+
+    // for (auto iter: QPDFDictItems(dictionary_obj))
+    // {
+    //     // iter.first is a string
+    //     // iter.second is a QPDFObjectHandle
+    // }
+
+  public:
+    QPDF_DLL
+    QPDFDictItems(QPDFObjectHandle& oh);
+
+    class iterator: public std::iterator<
+        std::bidirectional_iterator_tag,
+        std::pair<std::string, QPDFObjectHandle>>
+    {
+        friend class QPDFDictItems;
+      public:
+        QPDF_DLL
+        virtual ~iterator() = default;
+        QPDF_DLL
+        iterator& operator++();
+        QPDF_DLL
+        iterator operator++(int)
+        {
+            iterator t = *this;
+            ++(*this);
+            return t;
+        }
+        QPDF_DLL
+        iterator& operator--();
+        QPDF_DLL
+        iterator operator--(int)
+        {
+            iterator t = *this;
+            --(*this);
+            return t;
+        }
+        QPDF_DLL
+        reference operator*();
+        QPDF_DLL
+        pointer operator->();
+        QPDF_DLL
+        bool operator==(iterator const& other) const;
+        QPDF_DLL
+        bool operator!=(iterator const& other) const
+        {
+            return ! operator==(other);
+        }
+
+      private:
+        iterator(QPDFObjectHandle& oh, bool for_begin);
+        void updateIValue();
+
+        class Members
+        {
+            friend class QPDFDictItems::iterator;
+
+          public:
+            QPDF_DLL
+            ~Members() = default;
+
+          private:
+            Members(QPDFObjectHandle& oh, bool for_begin);
+            Members() = delete;
+            Members(Members const&) = delete;
+
+            QPDFObjectHandle& oh;
+            std::set<std::string> keys;
+            std::set<std::string>::iterator iter;
+            bool is_end;
+        };
+        PointerHolder<Members> m;
+        value_type ivalue;
+    };
+
+    QPDF_DLL
+    iterator begin();
+    QPDF_DLL
+    iterator end();
+
+  private:
+    QPDFObjectHandle& oh;
+};
+
+class QPDFArrayItems
+{
+    // This class allows C++-style iteration, including range-for
+    // iteration, around arrays. You can write
+
+    // for (auto iter: QPDFArrayItems(array_obj))
+    // {
+    //     // iter is a QPDFObjectHandle
+    // }
+
+  public:
+    QPDF_DLL
+    QPDFArrayItems(QPDFObjectHandle& oh);
+
+    class iterator: public std::iterator<
+        std::bidirectional_iterator_tag,
+        QPDFObjectHandle>
+    {
+        friend class QPDFArrayItems;
+      public:
+        QPDF_DLL
+        virtual ~iterator() = default;
+        QPDF_DLL
+        iterator& operator++();
+        QPDF_DLL
+        iterator operator++(int)
+        {
+            iterator t = *this;
+            ++(*this);
+            return t;
+        }
+        QPDF_DLL
+        iterator& operator--();
+        QPDF_DLL
+        iterator operator--(int)
+        {
+            iterator t = *this;
+            --(*this);
+            return t;
+        }
+        QPDF_DLL
+        reference operator*();
+        QPDF_DLL
+        pointer operator->();
+        QPDF_DLL
+        bool operator==(iterator const& other) const;
+        QPDF_DLL
+        bool operator!=(iterator const& other) const
+        {
+            return ! operator==(other);
+        }
+
+      private:
+        iterator(QPDFObjectHandle& oh, bool for_begin);
+        void updateIValue();
+
+        class Members
+        {
+            friend class QPDFArrayItems::iterator;
+
+          public:
+            QPDF_DLL
+            ~Members() = default;
+
+          private:
+            Members(QPDFObjectHandle& oh, bool for_begin);
+            Members() = delete;
+            Members(Members const&) = delete;
+
+            QPDFObjectHandle& oh;
+            int item_number;
+            bool is_end;
+        };
+        PointerHolder<Members> m;
+        value_type ivalue;
+    };
+
+    QPDF_DLL
+    iterator begin();
+    QPDF_DLL
+    iterator end();
+
+  private:
+    QPDFObjectHandle& oh;
+};
+
+
 #endif // QPDFOBJECTHANDLE_HH
diff --git a/libqpdf/QPDFObjectHandle.cc b/libqpdf/QPDFObjectHandle.cc
index 93f49792..288b5256 100644
--- a/libqpdf/QPDFObjectHandle.cc
+++ b/libqpdf/QPDFObjectHandle.cc
@@ -3079,3 +3079,180 @@ QPDFObjectHandle::warn(QPDF* qpdf, QPDFExc const& e)
         throw e;
     }
 }
+
+QPDFDictItems::QPDFDictItems(QPDFObjectHandle& oh) :
+    oh(oh)
+{
+}
+
+QPDFDictItems::iterator&
+QPDFDictItems::iterator::operator++()
+{
+    ++this->m->iter;
+    updateIValue();
+    return *this;
+}
+
+QPDFDictItems::iterator&
+QPDFDictItems::iterator::operator--()
+{
+    --this->m->iter;
+    updateIValue();
+    return *this;
+}
+
+QPDFDictItems::iterator::reference
+QPDFDictItems::iterator:: operator*()
+{
+    updateIValue();
+    return this->ivalue;
+}
+
+QPDFDictItems::iterator::pointer
+QPDFDictItems::iterator::operator->()
+{
+    updateIValue();
+    return &this->ivalue;
+}
+
+bool
+QPDFDictItems::iterator::operator==(iterator const& other) const
+{
+    if (this->m->is_end && other.m->is_end)
+    {
+        return true;
+    }
+    if (this->m->is_end || other.m->is_end)
+    {
+        return false;
+    }
+    return (this->ivalue.first == other.ivalue.first);
+}
+
+QPDFDictItems::iterator::iterator(QPDFObjectHandle& oh, bool for_begin) :
+    m(new Members(oh, for_begin))
+{
+    updateIValue();
+}
+
+void
+QPDFDictItems::iterator::updateIValue()
+{
+    this->m->is_end = (this->m->iter == this->m->keys.end());
+    if (this->m->is_end)
+    {
+        this->ivalue.first = "";
+        this->ivalue.second = QPDFObjectHandle();
+    }
+    else
+    {
+        this->ivalue.first = *(this->m->iter);
+        this->ivalue.second = this->m->oh.getKey(this->ivalue.first);
+    }
+}
+
+QPDFDictItems::iterator::Members::Members(
+    QPDFObjectHandle& oh, bool for_begin) :
+    oh(oh)
+{
+    this->keys = oh.getKeys();
+    this->iter = for_begin ? this->keys.begin() : this->keys.end();
+}
+
+QPDFDictItems::iterator
+QPDFDictItems::begin()
+{
+    return iterator(oh, true);
+}
+
+QPDFDictItems::iterator
+QPDFDictItems::end()
+{
+    return iterator(oh, false);
+}
+
+QPDFArrayItems::QPDFArrayItems(QPDFObjectHandle& oh) :
+    oh(oh)
+{
+}
+
+QPDFArrayItems::iterator&
+QPDFArrayItems::iterator::operator++()
+{
+    if (! this->m->is_end)
+    {
+        ++this->m->item_number;
+        updateIValue();
+    }
+    return *this;
+}
+
+QPDFArrayItems::iterator&
+QPDFArrayItems::iterator::operator--()
+{
+    if (this->m->item_number > 0)
+    {
+        --this->m->item_number;
+        updateIValue();
+    }
+    return *this;
+}
+
+QPDFArrayItems::iterator::reference
+QPDFArrayItems::iterator:: operator*()
+{
+    updateIValue();
+    return this->ivalue;
+}
+
+QPDFArrayItems::iterator::pointer
+QPDFArrayItems::iterator::operator->()
+{
+    updateIValue();
+    return &this->ivalue;
+}
+
+bool
+QPDFArrayItems::iterator::operator==(iterator const& other) const
+{
+    return (this->m->item_number == other.m->item_number);
+}
+
+QPDFArrayItems::iterator::iterator(QPDFObjectHandle& oh, bool for_begin) :
+    m(new Members(oh, for_begin))
+{
+    updateIValue();
+}
+
+void
+QPDFArrayItems::iterator::updateIValue()
+{
+    this->m->is_end = (this->m->item_number >= this->m->oh.getArrayNItems());
+    if (this->m->is_end)
+    {
+        this->ivalue = QPDFObjectHandle();
+    }
+    else
+    {
+        this->ivalue = this->m->oh.getArrayItem(this->m->item_number);
+    }
+}
+
+QPDFArrayItems::iterator::Members::Members(
+    QPDFObjectHandle& oh, bool for_begin) :
+    oh(oh)
+{
+    this->item_number = for_begin ? 0 : oh.getArrayNItems();
+}
+
+QPDFArrayItems::iterator
+QPDFArrayItems::begin()
+{
+    return iterator(oh, true);
+}
+
+QPDFArrayItems::iterator
+QPDFArrayItems::end()
+{
+    return iterator(oh, false);
+}
diff --git a/manual/qpdf-manual.xml b/manual/qpdf-manual.xml
index ccc80d07..9a62f1da 100644
--- a/manual/qpdf-manual.xml
+++ b/manual/qpdf-manual.xml
@@ -4849,6 +4849,17 @@ print "\n";
        <itemizedlist>
         <listitem>
          <para>
+          Add <classname>QPDFDictItems</classname> and
+          <classname>QPDFArrayItems</classname> wrappers around
+          <classname>QPDFObjectHandle</classname>, allowing C++-style
+          iteration, including range-for iteration, over dictionary
+          and array QPDFObjectHandles. See comments in
+          <filename>include/qpdf/QPDFObjectHandle.hh</filename> for
+          details.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
           Add <function>warn</function> to
           <classname>QPDF</classname>'s public API.
          </para>
diff --git a/qpdf/test_driver.cc b/qpdf/test_driver.cc
index 99c5cf53..6cb5fc80 100644
--- a/qpdf/test_driver.cc
+++ b/qpdf/test_driver.cc
@@ -347,30 +347,29 @@ void runtest(int n, char const* filename1, char const* arg2)
 	else if (qtest.isArray())
 	{
 	    QTC::TC("qpdf", "main QTest array");
-	    int nitems = qtest.getArrayNItems();
 	    std::cout << "/QTest is an array with "
-		      << nitems << " items" << std::endl;
-	    for (int i = 0; i < nitems; ++i)
+		      << qtest.getArrayNItems() << " items" << std::endl;
+            int i = 0;
+	    for (auto& iter: QPDFArrayItems(qtest))
 	    {
 		QTC::TC("qpdf", "main QTest array indirect",
-			qtest.getArrayItem(i).isIndirect() ? 1 : 0);
+			iter.isIndirect() ? 1 : 0);
 		std::cout << "  item " << i << " is "
-			  << (qtest.getArrayItem(i).isIndirect() ? "in" : "")
+			  << (iter.isIndirect() ? "in" : "")
 			  << "direct" << std::endl;
+                ++i;
 	    }
 	}
 	else if (qtest.isDictionary())
 	{
 	    QTC::TC("qpdf", "main QTest dictionary");
 	    std::cout << "/QTest is a dictionary" << std::endl;
-	    std::set<std::string> keys = qtest.getKeys();
-	    for (std::set<std::string>::iterator iter = keys.begin();
-		 iter != keys.end(); ++iter)
-	    {
+            for (auto& iter: QPDFDictItems(qtest))
+            {
 		QTC::TC("qpdf", "main QTest dictionary indirect",
-			(qtest.getKey(*iter).isIndirect() ? 1 : 0));
-		std::cout << "  " << *iter << " is "
-			  << (qtest.getKey(*iter).isIndirect() ? "in" : "")
+			iter.second.isIndirect() ? 1 : 0);
+		std::cout << "  " << iter.first << " is "
+			  << (iter.second.isIndirect() ? "in" : "")
 			  << "direct" << std::endl;
 	    }
 	}
@@ -1539,7 +1538,38 @@ void runtest(int n, char const* filename1, char const* arg2)
         QPDFObjectHandle integer = qtest.getKey("/Integer");
         QPDFObjectHandle null = QPDFObjectHandle::newNull();
         assert(array.isArray());
+        {
+            // Exercise iterators directly
+            QPDFArrayItems ai(array);
+            auto i = ai.begin();
+            assert(i->getName() == "/Item0");
+            auto& i_value = *i;
+            --i;
+            assert(i->getName() == "/Item0");
+            ++i;
+            ++i;
+            ++i;
+            assert(i == ai.end());
+            ++i;
+            assert(i == ai.end());
+            assert(! i_value.isInitialized());
+            --i;
+            assert(i_value.getName() == "/Item2");
+            assert(i->getName() == "/Item2");
+        }
         assert(dictionary.isDictionary());
+        {
+            // Exercise iterators directly
+            QPDFDictItems di(dictionary);
+            auto i = di.begin();
+            assert(i->first == "/Key1");
+            auto& i_value = *i;
+            assert(i->second.getName() == "/Value1");
+            ++i;
+            ++i;
+            assert(i == di.end());
+            assert(! i_value.second.isInitialized());
+        }
         assert("" == qtest.getStringValue());
         array.getArrayItem(-1).assertNull();
         array.getArrayItem(16059).assertNull();
@@ -1599,6 +1629,10 @@ void runtest(int n, char const* filename1, char const* arg2)
                (r1.lly > 3.39) && (r1.lly < 3.41) &&
                (r1.urx > 5.59) && (r1.urx < 5.61) &&
                (r1.ury > 7.79) && (r1.ury < 7.81));
+        QPDFObjectHandle uninitialized;
+        assert(! uninitialized.isInitialized());
+        assert(! uninitialized.isInteger());
+        assert(! uninitialized.isDictionary());
     }
     else if (n == 43)
     {