Support copying objects from another QPDF object

This includes QPDF::copyForeignObject and supporting foreign objects
as arguments to addPage*.

ChangeLog

@@ -1,3 +1,21 @@
+2012-07-11  Jay Berkenbilt  <ejb@ql.org>
+
+	* Added new APIs to copy objects from one QPDF to another.  This
+	includes letting QPDF::addPage() (and QPDF::addPageAt()) accept a
+	page object from another QPDF and adding
+	QPDF::copyForeignObject().  See QPDF.hh for details.
+
+	* Add method QPDFObjectHandle::getOwningQPDF() to return the QPDF
+	object associated with an indirect QPDFObjectHandle.
+
+	* Add convenience methods to QPDFObjectHandle: assertIndirect(),
+	isPageObject(), isPagesObject()
+
+	* Cache when QPDF::pushInheritedAttributesToPage() has been called
+	to avoid traversing the pages trees multiple times.  This state is
+	cleared by QPDF::updateAllPagesCache() and ignored by
+	QPDF::flattenPagesTree().
+
 2012-07-08  Jay Berkenbilt  <ejb@ql.org>
 
 	* Add QPDFObjectHandle::newReserved to create a reserved object

TODO

@@ -28,35 +28,11 @@ Next
    can only be used by one thread at a time, but multiple threads can
    simultaneously use separate objects.
 
+ * Write some documentation about the design of copyForeignObject.
 
-Soon
+ * copyForeignObject still to do:
-====
 
- * Provide an option to copy encryption parameters from another file.
+    - qpdf command
-   This would make it possible to decrypt a file, manually work with
-   it, and then re-encrypt it using the original encryption parameters
-   including a possibly unknown owner password.
-
- * See if I can support the new encryption formats mentioned in the
-   open bug on sourceforge.  Check other sourceforge bugs.
-
- * Splitting/merging concepts
-
-   newPDF() could create a PDF with just a trailer, no pages, and a
-   minimal info.  Then the page routines could be used to add pages to
-   it.
-
-   Starting with any pdf, you should be able to copy objects from
-   another pdf.  The copy should be smart about never traversing into
-   a /Page or /Pages.
-
-   We could provide a method of copying objects from one PDF into
-   another.  This would do whatever optimization is necessary (maybe
-   just optimizePagesTree) and then traverse the set of objects
-   specified to find all objects referenced by the set.  Each of those
-   would be copied over with a table mapping old ID to new ID.  This
-   would be done from bottom up most likely disallowing cycles or
-   handling them sanely.
 
       Command line could be something like
 
@@ -65,7 +41,8 @@ Soon
       The first file referenced would be the one whose other data would
       be preserved (like trailer, info, encryption, outlines, etc.).
       --new as first file would just use an empty file as the starting
-   point.
+      point.  Be explicit about whether outlines, etc., are handled.
+      They are not handled initially.
 
       Example: to grab pages 1-5 from file1 and 11-15 from file2
 
@@ -73,31 +50,32 @@ Soon
 
       To implement this, we would remove all pages from file1 except
       pages 1 through 5.  Then we would take pages 11 through 15 from
-   file2 and add them to a set for transfer.  This would end up
+      file2, copy them to the file, and add them as pages.
-   generating a list of indirect objects.  We would copy those objects
-   shallowly to the new PDF keeping track of the mapping and replacing
-   any indirect object keys as appropriate, much like QPDFWriter does.
 
-   When all the objects are registered, we would add those pages to
+    - document that makeIndirectObject doesn't handle foreign objects
-   the result.
+      automatically because copying a foreign object is a big enough
+      deal that it should be explicit.  However addPages* does handle
+      foreign page objects automatically.
 
-   This approach could work for both splitting and merging.  It's
+    - Test /Outlines and see whether there's any point in handling
-   possible it could be implemented now without any new APIs, but most
+      them in the API.  Maybe just copying them over works.  What
-   of the work should be doable by the library with only a small set
+      about command line tool?  Also think about page labels.
-   of additions.
 
-   newPDF()
+    - Tests through qpdf command line: copy pages from multiple PDFs
-   QPDFObjectCopier c(qpdf1, qpdf2)
+      starting with one PDF and also starting with empty.
-   QPDFObjectHandle obj = c.copyObject(<object from qpdf1>)
+
-     Without traversing pages, copies all indirect objects referenced
+ * (Hopefully) Provide an option to copy encryption parameters from
-     by <object from qpdf1> preserving referential integrity and
+   another file.  This would make it possible to decrypt a file,
-     returns an object handle in qpdf2 of the same object.  If called
+   manually work with it, and then re-encrypt it using the original
-     multiple times on the same object, retraverses in case there were
+   encryption parameters including a possibly unknown owner password.
-     changes.
+
+
+Soon
+====
+
+ * See if I can support the new encryption formats mentioned in the
+   open bug on sourceforge.  Check other sourceforge bugs.
 
-   QPDFObjectHandle obj = c.getMapping(<object from qpdf1>)
-     find the object in qpdf2 corresponding to the object from qpdf1.
-     Return the null object if none.
 
 General
 =======
@@ -110,23 +88,11 @@ General
  * Update qpdf docs about non-ascii passwords.  See thread from
    2010-12-07,08 for details.
 
- * Look at page splitting.  Subramanyam provided a test file; see
+ * Consider impact of article threads on page splitting/merging.
-   ../misc/article-threads.pdf.  Email Q-Count: 431864 from
+   Subramanyam provided a test file; see ../misc/article-threads.pdf.
-   2009-11-03.  See also "Splitting by Pages" below.
+   Email Q-Count: 431864 from 2009-11-03.  Other things to consider:
-
+   outlines, page labels, thumbnails, zones.  There are probably
- * Consider writing a PDF merge utility.  With 2.2, it would be
+   others.
-   possible to have a StreamDataProvider that would allow stream data
-   to be directly copied from one PDF file to another.  One possible
-   strategy would be to have a program that adds all the pages of one
-   file to the end of another file.  The basic
-   strategy would be to create a table that adds new streams to the
-   original file, mapping the new streams' obj/gen to a stream in the
-   file whose pages are being appended.  The StreamDataProvider, when
-   asked, could simply pipe the streams of the file being appended to
-   the provided pipeline and could copy the filter and decode
-   parameters from the original file.  Being able to do this requires
-   a lot of the same logic as being able to do splitting, so a general
-   split/merge program would be a great addition.
 
  * See whether it's possible to remove the call to
    flattenScalarReferences.  I can't easily figure out why I do it,
@@ -279,26 +245,3 @@ Index: QPDFWriter.cc
 
  * From a suggestion in bug 3152169, consisder having an option to
    re-encode inline images with an ASCII encoding.
-
-
-Splitting by Pages
-==================
-
-Although qpdf does not currently support splitting a file into pages,
-the work done for linearization covers almost all the work.  To do
-page splitting.  If this functionality is needed, study
-obj_user_to_objects and object_to_obj_users created in
-QPDF_optimization for ideas.  It's quite possible that the information
-computed by calculateLinearizationData is actually sufficient to do
-page splitting in many circumstances.  That code knows which objects
-are used by which pages, though it doesn't do anything page-specific
-with outlines, thumbnails, page labels, or anything else.
-
-Another approach would be to traverse only pages that are being output
-taking care not to traverse into the pages tree, and then to fabricate
-a new pages tree.
-
-Either way, care must be taken to handle other things such as
-outlines, page labels, thumbnails, threads, zones, etc. in a sensible
-way.  This may include simply omitting information other than page
-content.

include/qpdf/QPDF.hh

@@ -190,6 +190,28 @@ class QPDF
     replaceReserved(QPDFObjectHandle reserved,
                     QPDFObjectHandle replacement);
 
+    // Copy an object from another QPDF to this one.  The return value
+    // is an indirect reference to the copied object in this file.
+    // This method is intended to be used to copy non-page objects and
+    // will not copy page objects.  To copy page objects, pass the
+    // foreign page object directly to addPage (or addPageAt).  If you
+    // copy objects that contain references to pages, you should copy
+    // the pages first using addPage(At).  Otherwise references to the
+    // pages that have not been copied will be replaced with nulls.
+
+    // When copying objects with this method, object structure will be
+    // preserved, so all indirectly referenced indirect objects will
+    // be copied as well.  This includes any circular references that
+    // may exist.  The QPDF object keeps a record of what has already
+    // been copied, so shared objects will not be copied multiple
+    // times.  This also means that if you mutate an object that has
+    // already been copied and try to copy it again, it won't work
+    // since the modified object will not be recopied.  Therefore, you
+    // should do all mutation on the original file that you are going
+    // to do before you start copying its objects to a new file.
+    QPDF_DLL
+    QPDFObjectHandle copyForeignObject(QPDFObjectHandle foreign);
+
     // Encryption support
 
     enum encryption_method_e { e_none, e_unknown, e_rc4, e_aes };
@@ -380,7 +402,10 @@ class QPDF
     // modify /Pages structures directly, you must call this method
     // afterwards.  This method updates the internal list of pages, so
     // after calling this method, any previous references returned by
-    // getAllPages() will be valid again.
+    // getAllPages() will be valid again.  It also resets any state
+    // about having pushed inherited attributes in /Pages objects down
+    // to the pages, so if you add any inheritable attributes to a
+    // /Pages object, you should also call this method.
     QPDF_DLL
     void updateAllPagesCache();
 
@@ -389,11 +414,19 @@ class QPDF
     // resolved by explicitly setting the values in each /Page.
     void pushInheritedAttributesToPage();
 
-    // Add new page at the beginning or the end of the current pdf
+    // Add new page at the beginning or the end of the current pdf.
+    // The newpage parameter may be either a direct object, an
+    // indirect object from this QPDF, or an indirect object from
+    // another QPDF.  If it is a direct object, it will be made
+    // indirect.  If it is an indirect object from another QPDF, this
+    // method will call pushInheritedAttributesToPage on the other
+    // file and then copy the page to this QPDF using the same
+    // underlying code as copyForeignObject.
     QPDF_DLL
     void addPage(QPDFObjectHandle newpage, bool first);
 
-    // Add new page before or after refpage
+    // Add new page before or after refpage.  See comments for addPage
+    // for details about what newpage should be.
     QPDF_DLL
     void addPageAt(QPDFObjectHandle newpage, bool before,
                    QPDFObjectHandle refpage);
@@ -542,6 +575,29 @@ class QPDF
 	qpdf_offset_t end_after_space;
     };
 
+    class ObjCopier
+    {
+      public:
+        std::map<ObjGen, QPDFObjectHandle> object_map;
+        std::vector<QPDFObjectHandle> to_copy;
+        std::set<ObjGen> visiting;
+    };
+
+    class CopiedStreamDataProvider: public QPDFObjectHandle::StreamDataProvider
+    {
+      public:
+        virtual ~CopiedStreamDataProvider()
+        {
+        }
+	virtual void provideStreamData(int objid, int generation,
+				       Pipeline* pipeline);
+        void registerForeignStream(ObjGen const& local_og,
+                                   QPDFObjectHandle foreign_stream);
+
+      private:
+        std::map<ObjGen, QPDFObjectHandle> foreign_streams;
+    };
+
     void parse(char const* password);
     void warn(QPDFExc const& e);
     void setTrailer(QPDFObjectHandle obj);
@@ -602,6 +658,14 @@ class QPDF
 	QPDFObjectHandle& stream_dict,
 	std::vector<PointerHolder<Pipeline> >& heap);
 
+    // Methods to support object copying
+    QPDFObjectHandle copyForeignObject(
+        QPDFObjectHandle foreign, bool allow_page);
+    void reserveObjects(QPDFObjectHandle foreign, ObjCopier& obj_copier,
+                        bool top);
+    QPDFObjectHandle replaceForeignIndirectObjects(
+        QPDFObjectHandle foreign, ObjCopier& obj_copier, bool top);
+
     // Linearization Hint table structures.
     // Naming conventions:
 
@@ -960,7 +1024,12 @@ class QPDF
     QPDFObjectHandle trailer;
     std::vector<QPDFObjectHandle> all_pages;
     std::map<ObjGen, int> pageobj_to_pages_pos;
+    bool pushed_inherited_attributes_to_pages;
     std::vector<QPDFExc> warnings;
+    std::map<QPDF*, ObjCopier> object_copiers;
+    PointerHolder<QPDFObjectHandle::StreamDataProvider> copied_streams;
+    // copied_stream_data_provider is owned by copied_streams
+    CopiedStreamDataProvider* copied_stream_data_provider;
 
     // Linearization data
     qpdf_offset_t first_xref_item_offset; // actual value from file

include/qpdf/QPDFObjectHandle.hh

@@ -222,6 +222,11 @@ class QPDFObjectHandle
     QPDF_DLL
     bool isOrHasName(std::string const&);
 
+    // Return the QPDF object that owns an indirect object.  Returns
+    // null for a direct object.
+    QPDF_DLL
+    QPDF* getOwningQPDF();
+
     // Create a shallow copy of an object as a direct object.  Since
     // this is a shallow copy, for dictionaries and arrays, any keys
     // or items that were indirect objects will still be indirect
@@ -453,10 +458,17 @@ class QPDFObjectHandle
     QPDF_DLL
     void assertReserved();
 
+    QPDF_DLL
+    void assertIndirect();
     QPDF_DLL
     void assertScalar();
     QPDF_DLL
     void assertNumber();
+
+    QPDF_DLL
+    bool isPageObject();
+    QPDF_DLL
+    bool isPagesObject();
     QPDF_DLL
     void assertPageObject();
 

libqpdf/QPDF.cc

@@ -348,6 +348,23 @@ QPDF::ObjGen::operator<(ObjGen const& rhs) const
 	    ((this->obj == rhs.obj) && (this->gen < rhs.gen)));
 }
 
+void
+QPDF::CopiedStreamDataProvider::provideStreamData(
+    int objid, int generation, Pipeline* pipeline)
+{
+    QPDFObjectHandle foreign_stream =
+        this->foreign_streams[ObjGen(objid, generation)];
+    foreign_stream.pipeStreamData(pipeline, false, false, false);
+}
+
+void
+QPDF::CopiedStreamDataProvider::registerForeignStream(
+    ObjGen const& local_og, QPDFObjectHandle foreign_stream)
+{
+    this->foreign_streams[local_og] = foreign_stream;
+}
+
+
 std::string const&
 QPDF::QPDFVersion()
 {
@@ -369,6 +386,8 @@ QPDF::QPDF() :
     cf_file(e_none),
     cached_key_objid(0),
     cached_key_generation(0),
+    pushed_inherited_attributes_to_pages(false),
+    copied_stream_data_provider(0),
     first_xref_item_offset(0),
     uncompressed_after_compressed(false)
 {
@@ -2067,6 +2086,244 @@ QPDF::replaceReserved(QPDFObjectHandle reserved,
                   replacement);
 }
 
+QPDFObjectHandle
+QPDF::copyForeignObject(QPDFObjectHandle foreign)
+{
+    return copyForeignObject(foreign, false);
+}
+
+QPDFObjectHandle
+QPDF::copyForeignObject(QPDFObjectHandle foreign, bool allow_page)
+{
+    if (! foreign.isIndirect())
+    {
+        QTC::TC("qpdf", "QPDF copyForeign direct");
+	throw std::logic_error(
+	    "QPDF::copyForeign called with direct object handle");
+    }
+    QPDF* other = foreign.getOwningQPDF();
+    if (other == this)
+    {
+        QTC::TC("qpdf", "QPDF copyForeign not foreign");
+        throw std::logic_error(
+            "QPDF::copyForeign called with object from this QPDF");
+    }
+
+    ObjCopier& obj_copier = this->object_copiers[other];
+    if (! obj_copier.visiting.empty())
+    {
+        throw std::logic_error("obj_copier.visiting is not empty"
+                               " at the beginning of copyForeignObject");
+    }
+
+    // Make sure we have an object in this file for every referenced
+    // object in the old file.  obj_copier.object_map maps foreign
+    // ObjGen to local objects.  For everything new that we have to
+    // copy, the local object will be a reservation, unless it is a
+    // stream, in which case the local object will already be a
+    // stream.
+    reserveObjects(foreign, obj_copier, true);
+
+    if (! obj_copier.visiting.empty())
+    {
+        throw std::logic_error("obj_copier.visiting is not empty"
+                               " after reserving objects");
+    }
+
+    // Copy any new objects and replace the reservations.
+    for (std::vector<QPDFObjectHandle>::iterator iter =
+             obj_copier.to_copy.begin();
+         iter != obj_copier.to_copy.end(); ++iter)
+    {
+        QPDFObjectHandle& to_copy = *iter;
+        QPDFObjectHandle copy =
+            replaceForeignIndirectObjects(to_copy, obj_copier, true);
+        if (! to_copy.isStream())
+        {
+            ObjGen og(to_copy.getObjectID(), to_copy.getGeneration());
+            replaceReserved(obj_copier.object_map[og], copy);
+        }
+    }
+    obj_copier.to_copy.clear();
+
+    return obj_copier.object_map[ObjGen(foreign.getObjectID(),
+                                        foreign.getGeneration())];
+}
+
+void
+QPDF::reserveObjects(QPDFObjectHandle foreign, ObjCopier& obj_copier,
+                     bool top)
+{
+    if (foreign.isReserved())
+    {
+        throw std::logic_error(
+            "QPDF: attempting to copy a foreign reserved object");
+    }
+
+    if (foreign.isPagesObject())
+    {
+        QTC::TC("qpdf", "QPDF not copying pages object");
+        return;
+    }
+
+    if ((! top) && foreign.isPageObject())
+    {
+        QTC::TC("qpdf", "QPDF not crossing page boundary");
+        return;
+    }
+
+    if (foreign.isIndirect())
+    {
+        ObjGen foreign_og(foreign.getObjectID(), foreign.getGeneration());
+        if (obj_copier.visiting.find(foreign_og) != obj_copier.visiting.end())
+        {
+            QTC::TC("qpdf", "QPDF loop reserving objects");
+            return;
+        }
+        QTC::TC("qpdf", "QPDF copy indirect");
+        obj_copier.visiting.insert(foreign_og);
+        std::map<ObjGen, QPDFObjectHandle>::iterator mapping =
+            obj_copier.object_map.find(foreign_og);
+        if (mapping == obj_copier.object_map.end())
+        {
+            obj_copier.to_copy.push_back(foreign);
+            QPDFObjectHandle reservation;
+            if (foreign.isStream())
+            {
+                reservation = QPDFObjectHandle::newStream(this);
+            }
+            else
+            {
+                reservation = QPDFObjectHandle::newReserved(this);
+            }
+            obj_copier.object_map[foreign_og] = reservation;
+        }
+    }
+
+    if (foreign.isArray())
+    {
+        QTC::TC("qpdf", "QPDF reserve array");
+	int n = foreign.getArrayNItems();
+	for (int i = 0; i < n; ++i)
+	{
+            reserveObjects(foreign.getArrayItem(i), obj_copier, false);
+	}
+    }
+    else if (foreign.isDictionary())
+    {
+        QTC::TC("qpdf", "QPDF reserve dictionary");
+	std::set<std::string> keys = foreign.getKeys();
+	for (std::set<std::string>::iterator iter = keys.begin();
+	     iter != keys.end(); ++iter)
+	{
+            reserveObjects(foreign.getKey(*iter), obj_copier, false);
+	}
+    }
+    else if (foreign.isStream())
+    {
+        QTC::TC("qpdf", "QPDF reserve stream");
+        reserveObjects(foreign.getDict(), obj_copier, false);
+    }
+
+    if (foreign.isIndirect())
+    {
+        ObjGen foreign_og(foreign.getObjectID(), foreign.getGeneration());
+        obj_copier.visiting.erase(foreign_og);
+    }
+}
+
+QPDFObjectHandle
+QPDF::replaceForeignIndirectObjects(
+    QPDFObjectHandle foreign, ObjCopier& obj_copier, bool top)
+{
+    QPDFObjectHandle result;
+    if ((! top) && foreign.isIndirect())
+    {
+        QTC::TC("qpdf", "QPDF replace indirect");
+        ObjGen foreign_og(foreign.getObjectID(), foreign.getGeneration());
+        std::map<ObjGen, QPDFObjectHandle>::iterator mapping =
+            obj_copier.object_map.find(foreign_og);
+        if (mapping == obj_copier.object_map.end())
+        {
+            // This case would occur if this is a reference to a Page
+            // or Pages object that we didn't traverse into.
+            QTC::TC("qpdf", "QPDF replace foreign indirect with null");
+            result = QPDFObjectHandle::newNull();
+        }
+        else
+        {
+            result = obj_copier.object_map[foreign_og];
+        }
+    }
+    else if (foreign.isArray())
+    {
+        QTC::TC("qpdf", "QPDF replace array");
+        result = QPDFObjectHandle::newArray();
+	int n = foreign.getArrayNItems();
+	for (int i = 0; i < n; ++i)
+	{
+            result.appendItem(
+                replaceForeignIndirectObjects(
+                    foreign.getArrayItem(i), obj_copier, false));
+	}
+    }
+    else if (foreign.isDictionary())
+    {
+        QTC::TC("qpdf", "QPDF replace dictionary");
+        result = QPDFObjectHandle::newDictionary();
+	std::set<std::string> keys = foreign.getKeys();
+	for (std::set<std::string>::iterator iter = keys.begin();
+	     iter != keys.end(); ++iter)
+	{
+            result.replaceKey(
+                *iter,
+                replaceForeignIndirectObjects(
+                    foreign.getKey(*iter), obj_copier, false));
+	}
+    }
+    else if (foreign.isStream())
+    {
+        QTC::TC("qpdf", "QPDF replace stream");
+        ObjGen foreign_og(foreign.getObjectID(), foreign.getGeneration());
+        result = obj_copier.object_map[foreign_og];
+        result.assertStream();
+        QPDFObjectHandle dict = result.getDict();
+        QPDFObjectHandle old_dict = foreign.getDict();
+        std::set<std::string> keys = old_dict.getKeys();
+        for (std::set<std::string>::iterator iter = keys.begin();
+	     iter != keys.end(); ++iter)
+	{
+            dict.replaceKey(
+                *iter,
+                replaceForeignIndirectObjects(
+                    old_dict.getKey(*iter), obj_copier, false));
+	}
+        if (this->copied_stream_data_provider == 0)
+        {
+            this->copied_stream_data_provider = new CopiedStreamDataProvider();
+            this->copied_streams = this->copied_stream_data_provider;
+        }
+        ObjGen local_og(result.getObjectID(), result.getGeneration());
+        this->copied_stream_data_provider->registerForeignStream(
+            local_og, foreign);
+        result.replaceStreamData(this->copied_streams,
+                                 dict.getKey("/Filter"),
+                                 dict.getKey("/DecodeParms"));
+    }
+    else
+    {
+        foreign.assertScalar();
+        result = foreign;
+        result.makeDirect();
+    }
+
+    if (top && (! result.isStream()) && result.isIndirect())
+    {
+        throw std::logic_error("replacement for foreign object is indirect");
+    }
+
+    return result;
+}
 
 void
 QPDF::swapObjects(int objid1, int generation1, int objid2, int generation2)

libqpdf/QPDFObjectHandle.cc

@@ -355,6 +355,14 @@ QPDFObjectHandle::isOrHasName(std::string const& value)
     return false;
 }
 
+// Indirect object accessors
+QPDF*
+QPDFObjectHandle::getOwningQPDF()
+{
+    // Will be null for direct objects
+    return this->qpdf;
+}
+
 // Dictionary mutators
 
 void
@@ -784,6 +792,7 @@ QPDFObjectHandle::makeDirectInternal(std::set<int>& visited)
     }
 
     dereference();
+    this->qpdf = 0;
     this->objid = 0;
     this->generation = 0;
 
@@ -945,6 +954,16 @@ QPDFObjectHandle::assertReserved()
     assertType("Reserved", isReserved());
 }
 
+void
+QPDFObjectHandle::assertIndirect()
+{
+    if (! isIndirect())
+    {
+	throw std::logic_error(
+            "operation for indirect object attempted on direct object");
+    }
+}
+
 void
 QPDFObjectHandle::assertScalar()
 {
@@ -957,11 +976,24 @@ QPDFObjectHandle::assertNumber()
     assertType("Number", isNumber());
 }
 
+bool
+QPDFObjectHandle::isPageObject()
+{
+    return (this->isDictionary() && this->hasKey("/Type") &&
+            (this->getKey("/Type").getName() == "/Page"));
+}
+
+bool
+QPDFObjectHandle::isPagesObject()
+{
+    return (this->isDictionary() && this->hasKey("/Type") &&
+            (this->getKey("/Type").getName() == "/Pages"));
+}
+
 void
 QPDFObjectHandle::assertPageObject()
 {
-    if (! (this->isDictionary() && this->hasKey("/Type") &&
+    if (! isPageObject())
-	   (this->getKey("/Type").getName() == "/Page")))
     {
 	throw std::logic_error("page operation called on non-Page object");
     }

libqpdf/QPDF_optimization.cc

@@ -232,6 +232,14 @@ QPDF::pushInheritedAttributesToPage(bool allow_changes, bool warn_skipped_keys)
     // Traverse pages tree pushing all inherited resources down to the
     // page level.
 
+    // The record of whether we've done this is cleared by
+    // updateAllPagesCache().  If we're warning for skipped keys,
+    // re-traverse unconditionally.
+    if (this->pushed_inherited_attributes_to_pages && (! warn_skipped_keys))
+    {
+        return;
+    }
+
     // key_ancestors is a mapping of page attribute keys to a stack of
     // Pages nodes that contain values for them.
     std::map<std::string, std::vector<QPDFObjectHandle> > key_ancestors;
@@ -240,6 +248,7 @@ QPDF::pushInheritedAttributesToPage(bool allow_changes, bool warn_skipped_keys)
         this->trailer.getKey("/Root").getKey("/Pages"),
         key_ancestors, this->all_pages, allow_changes, warn_skipped_keys);
     assert(key_ancestors.empty());
+    this->pushed_inherited_attributes_to_pages = true;
 }
 
 void

libqpdf/QPDF_pages.cc

@@ -89,6 +89,7 @@ QPDF::updateAllPagesCache()
     QTC::TC("qpdf", "QPDF updateAllPagesCache");
     this->all_pages.clear();
     this->pageobj_to_pages_pos.clear();
+    this->pushed_inherited_attributes_to_pages = false;
     getAllPages();
 }
 
@@ -161,6 +162,12 @@ QPDF::insertPage(QPDFObjectHandle newpage, int pos)
         QTC::TC("qpdf", "QPDF insert non-indirect page");
         newpage = this->makeIndirectObject(newpage);
     }
+    else if (newpage.getOwningQPDF() != this)
+    {
+        QTC::TC("qpdf", "QPDF insert foreign page");
+        newpage.getOwningQPDF()->pushInheritedAttributesToPage();
+        newpage = this->copyForeignObject(newpage, true);
+    }
     else
     {
         QTC::TC("qpdf", "QPDF insert indirect page");

qpdf/qpdf.testcov

@@ -218,3 +218,18 @@ QPDF unknown key not inherited 0
 QPDF_Stream provider length not provided 0
 QPDF_Stream unknown stream length 0
 QPDF replaceReserved 0
+QPDF copyForeign direct 0
+QPDF copyForeign not foreign 0
+QPDF copy indirect 0
+QPDF loop reserving objects 0
+QPDF replace indirect 0
+QPDF replace array 0
+QPDF replace dictionary 0
+QPDF replace stream 0
+QPDF reserve array 0
+QPDF reserve dictionary 0
+QPDF reserve stream 0
+QPDF not crossing page boundary 0
+QPDF replace foreign indirect with null 0
+QPDF not copying pages object 0
+QPDF insert foreign page 0

qpdf/qtest/qpdf.test

@@ -379,6 +379,27 @@ $td->runtest("check output",
 	     {$td->FILE => "a.pdf"},
 	     {$td->FILE => "from-scratch-0.pdf"});
 # ----------
+$td->notify("--- Copy Foreign Objects ---");
+$n_tests += 7;
+
+foreach my $d ([25, 1], [26, 2], [27, 3])
+{
+    my ($testn, $outn) = @$d;
+    $td->runtest("copy objects $outn",
+                 {$td->COMMAND => "test_driver $testn" .
+                      " copy-foreign-objects-in.pdf"},
+                 {$td->STRING => "test $testn done\n", $td->EXIT_STATUS => 0},
+                 $td->NORMALIZE_NEWLINES);
+    $td->runtest("check output",
+                 {$td->FILE => "a.pdf"},
+                 {$td->FILE => "copy-foreign-objects-out$outn.pdf"});
+}
+$td->runtest("copy objects error",
+             {$td->COMMAND => "test_driver 28 copy-foreign-objects-in.pdf"},
+             {$td->FILE => "copy-foreign-objects-errors.out",
+              $td->EXIT_STATUS => 0},
+             $td->NORMALIZE_NEWLINES);
+# ----------
 $td->notify("--- Error Condition Tests ---");
 # $n_tests incremented after initialization of badfiles below.
 

qpdf/qtest/qpdf/copy-foreign-objects-errors.out

@@ -0,0 +1,3 @@
+logic error: QPDF::copyForeign called with object from this QPDF
+logic error: QPDF::copyForeign called with direct object handle
+test 28 done

qpdf/qtest/qpdf/copy-foreign-objects-in.pdf

@@ -0,0 +1,335 @@
+%PDF-1.3
+%¿÷¢þ
+%QDF-1.0
+
+% This test file is specifically crafted for testing copyForeignObject
+% and also for testing addPage when called with a page from another
+% file.
+
+% The /QTest key in trailer has pointers to several indirect objects:
+% O1, O2, O3 where O1 is an array that contains a dictionary that has
+% a key that points to O2, O2 is a dictionary that contains an array
+% that points to O1, and O3 is a page object that inherits some
+% resource from its parent /Pages and also points to some other page.
+% O1 also points to a stream whose dictionary has a key that points to
+% another stream whose dictionary points back to the first stream.
+
+1 0 obj
+<<
+  /Pages 2 0 R
+  /Type /Catalog
+>>
+endobj
+
+2 0 obj
+<<
+  /Count 5
+  /Kids [
+    3 0 R
+    4 0 R
+    5 0 R
+    6 0 R
+    7 0 R
+  ]
+  /Rotate 180
+  /Type /Pages
+>>
+endobj
+
+%% Page 1
+3 0 obj
+<<
+  /Contents 8 0 R
+  /MediaBox [
+    0
+    0
+    612
+    792
+  ]
+  /Parent 2 0 R
+  /Resources <<
+    /Font <<
+      /F1 10 0 R
+    >>
+    /ProcSet [
+      /PDF
+      /Text
+    ]
+  >>
+  /Type /Page
+>>
+endobj
+
+%% Page 2
+4 0 obj
+<<
+  /Contents 11 0 R
+  /MediaBox [
+    0
+    0
+    612
+    792
+  ]
+  /Parent 2 0 R
+  /Resources <<
+    /Font <<
+      /F1 10 0 R
+    >>
+    /ProcSet [
+      /PDF
+      /Text
+    ]
+  >>
+  /Type /Page
+>>
+endobj
+
+%% Page 3, object O3
+5 0 obj
+<<
+  /This-is-O3 true
+  /Contents 13 0 R
+  /MediaBox [
+    0
+    0
+    612
+    792
+  ]
+  /Parent 2 0 R
+  /Resources <<
+    /Font <<
+      /F1 10 0 R
+    >>
+    /ProcSet [
+      /PDF
+      /Text
+    ]
+  >>
+  /OtherPage 6 0 R
+  /Type /Page
+>>
+endobj
+
+%% Page 4
+6 0 obj
+<<
+  /This-is-O3-other-page true
+  /Contents 15 0 R
+  /MediaBox [
+    0
+    0
+    612
+    792
+  ]
+  /Parent 2 0 R
+  /Resources <<
+    /Font <<
+      /F1 10 0 R
+    >>
+    /ProcSet [
+      /PDF
+      /Text
+    ]
+  >>
+  /Type /Page
+>>
+endobj
+
+%% Page 5
+7 0 obj
+<<
+  /Contents 17 0 R
+  /MediaBox [
+    0
+    0
+    612
+    792
+  ]
+  /Parent 2 0 R
+  /Resources <<
+    /Font <<
+      /F1 10 0 R
+    >>
+    /ProcSet [
+      /PDF
+      /Text
+    ]
+  >>
+  /Type /Page
+>>
+endobj
+
+%% Contents for page 1
+8 0 obj
+<<
+  /Length 9 0 R
+>>
+stream
+BT /F1 15 Tf 72 720 Td (Original page 0) Tj ET
+endstream
+endobj
+
+9 0 obj
+47
+endobj
+
+10 0 obj
+<<
+  /BaseFont /Times-Roman
+  /Encoding /WinAnsiEncoding
+  /Subtype /Type1
+  /Type /Font
+>>
+endobj
+
+%% Contents for page 2
+11 0 obj
+<<
+  /Length 12 0 R
+>>
+stream
+BT /F1 15 Tf 72 720 Td (Original page 1) Tj ET
+endstream
+endobj
+
+12 0 obj
+47
+endobj
+
+%% Contents for page 3
+13 0 obj
+<<
+  /Length 14 0 R
+>>
+stream
+BT /F1 15 Tf 72 720 Td (Original page 2) Tj ET
+endstream
+endobj
+
+14 0 obj
+47
+endobj
+
+%% Contents for page 4
+15 0 obj
+<<
+  /Length 16 0 R
+>>
+stream
+BT /F1 15 Tf 72 720 Td (Original page 3) Tj ET
+endstream
+endobj
+
+16 0 obj
+47
+endobj
+
+%% Contents for page 5
+17 0 obj
+<<
+  /Length 18 0 R
+>>
+stream
+BT /F1 15 Tf 72 720 Td (Original page 4) Tj ET
+endstream
+endobj
+
+18 0 obj
+47
+endobj
+
+% O1
+19 0 obj
+[
+  /This-is-O1
+  /potato
+  << /O2 [3.14159 << /O2 20 0 R >> 2.17828 ] >>
+  /salad
+  /O2 20 0 R
+  /Stream1 21 0 R
+]
+endobj
+
+% O2
+20 0 obj
+<<
+  /This-is-O2 true
+  /K1 [2.236 /O1 19 0 R 1.732]
+  /O1 19 0 R
+>>
+endobj
+
+% stream1
+21 0 obj
+<<
+  /This-is-Stream1 true
+  /Length 22 0 R
+  /Stream2 23 0 R
+>>
+stream
+This is stream 1.
+endstream
+endobj
+
+22 0 obj
+18
+endobj
+
+% stream2
+23 0 obj
+<<
+  /This-is-Stream2 true
+  /Length 24 0 R
+  /Stream1 21 0 R
+>>
+stream
+This is stream 2.
+endstream
+endobj
+
+24 0 obj
+18
+endobj
+
+% QTest
+25 0 obj
+<< /This-is-QTest true /O1 19 0 R /O2 20 0 R /O3 5 0 R >>
+endobj
+
+xref
+0 26
+0000000000 65535 f 
+0000000655 00000 n 
+0000000709 00000 n 
+0000000845 00000 n 
+0000001073 00000 n 
+0000001313 00000 n 
+0000001580 00000 n 
+0000001839 00000 n 
+0000002081 00000 n 
+0000002183 00000 n 
+0000002202 00000 n 
+0000002334 00000 n 
+0000002438 00000 n 
+0000002481 00000 n 
+0000002585 00000 n 
+0000002628 00000 n 
+0000002732 00000 n 
+0000002775 00000 n 
+0000002879 00000 n 
+0000002904 00000 n 
+0000003042 00000 n 
+0000003138 00000 n 
+0000003255 00000 n 
+0000003285 00000 n 
+0000003402 00000 n 
+0000003430 00000 n 
+trailer <<
+  /Root 1 0 R
+  /Size 26
+  /QTest 25 0 R
+  /ID [<d15f7aca3be584a96c1c94adb0931e71><9adb6b2fdb22e857340f7103917b16e4>]
+>>
+startxref
+3505
+%%EOF

qpdf/qtest/qpdf/copy-foreign-objects-out1.pdf

@@ -0,0 +1,66 @@
+%PDF-1.3
+%¿÷¢þ
+1 0 obj
+<< /Pages 3 0 R /Type /Catalog >>
+endobj
+2 0 obj
+<< /O1 4 0 R /O2 5 0 R /This-is-QTest true >>
+endobj
+3 0 obj
+<< /Count 1 /Kids [ 6 0 R ] /Type /Pages >>
+endobj
+4 0 obj
+[ /This-is-O1 /potato << /O2 [ 3.14159 << /O2 5 0 R >> 2.17828 ] >> /salad /O2 5 0 R /Stream1 7 0 R ]
+endobj
+5 0 obj
+<< /K1 [ 2.236 /O1 4 0 R 1.732 ] /O1 4 0 R /This-is-O2 true >>
+endobj
+6 0 obj
+<< /Contents 8 0 R /MediaBox [ 0 0 612 792 ] /Parent 3 0 R /Resources << /Font << /F1 9 0 R >> /ProcSet 10 0 R >> /Type /Page >>
+endobj
+7 0 obj
+<< /Stream2 11 0 R /This-is-Stream1 true /Length 18 >>
+stream
+This is stream 1.
+endstream
+endobj
+8 0 obj
+<< /Length 44 >>
+stream
+BT
+  /F1 24 Tf
+  72 720 Td
+  (Potato) Tj
+ET
+endstream
+endobj
+9 0 obj
+<< /BaseFont /Helvetica /Encoding /WinAnsiEncoding /Name /F1 /Subtype /Type1 /Type /Font >>
+endobj
+10 0 obj
+[ /PDF /Text ]
+endobj
+11 0 obj
+<< /Stream1 7 0 R /This-is-Stream2 true /Length 18 >>
+stream
+This is stream 2.
+endstream
+endobj
+xref
+0 12
+0000000000 65535 f 
+0000000015 00000 n 
+0000000064 00000 n 
+0000000125 00000 n 
+0000000184 00000 n 
+0000000301 00000 n 
+0000000379 00000 n 
+0000000523 00000 n 
+0000000628 00000 n 
+0000000721 00000 n 
+0000000828 00000 n 
+0000000859 00000 n 
+trailer << /QTest 2 0 R /Root 1 0 R /Size 12 /ID [<31415926535897932384626433832795><31415926535897932384626433832795>] >>
+startxref
+964
+%%EOF

qpdf/qtest/qpdf/copy-foreign-objects-out2.pdf

@@ -0,0 +1,81 @@
+%PDF-1.3
+%¿÷¢þ
+1 0 obj
+<< /Pages 3 0 R /Type /Catalog >>
+endobj
+2 0 obj
+<< /O1 4 0 R /O2 5 0 R /O3 6 0 R /This-is-QTest true >>
+endobj
+3 0 obj
+<< /Count 2 /Kids [ 7 0 R 6 0 R ] /Type /Pages >>
+endobj
+4 0 obj
+[ /This-is-O1 /potato << /O2 [ 3.14159 << /O2 5 0 R >> 2.17828 ] >> /salad /O2 5 0 R /Stream1 8 0 R ]
+endobj
+5 0 obj
+<< /K1 [ 2.236 /O1 4 0 R 1.732 ] /O1 4 0 R /This-is-O2 true >>
+endobj
+6 0 obj
+<< /Contents 9 0 R /MediaBox [ 0 0 612 792 ] /Parent 3 0 R /Resources << /Font << /F1 10 0 R >> /ProcSet [ /PDF /Text ] >> /Rotate 180 /This-is-O3 true /Type /Page >>
+endobj
+7 0 obj
+<< /Contents 11 0 R /MediaBox [ 0 0 612 792 ] /Parent 3 0 R /Resources << /Font << /F1 12 0 R >> /ProcSet 13 0 R >> /Type /Page >>
+endobj
+8 0 obj
+<< /Stream2 14 0 R /This-is-Stream1 true /Length 18 >>
+stream
+This is stream 1.
+endstream
+endobj
+9 0 obj
+<< /Length 47 >>
+stream
+BT /F1 15 Tf 72 720 Td (Original page 2) Tj ET
+endstream
+endobj
+10 0 obj
+<< /BaseFont /Times-Roman /Encoding /WinAnsiEncoding /Subtype /Type1 /Type /Font >>
+endobj
+11 0 obj
+<< /Length 44 >>
+stream
+BT
+  /F1 24 Tf
+  72 720 Td
+  (Potato) Tj
+ET
+endstream
+endobj
+12 0 obj
+<< /BaseFont /Helvetica /Encoding /WinAnsiEncoding /Name /F1 /Subtype /Type1 /Type /Font >>
+endobj
+13 0 obj
+[ /PDF /Text ]
+endobj
+14 0 obj
+<< /Stream1 8 0 R /This-is-Stream2 true /Length 18 >>
+stream
+This is stream 2.
+endstream
+endobj
+xref
+0 15
+0000000000 65535 f 
+0000000015 00000 n 
+0000000064 00000 n 
+0000000135 00000 n 
+0000000200 00000 n 
+0000000317 00000 n 
+0000000395 00000 n 
+0000000577 00000 n 
+0000000723 00000 n 
+0000000828 00000 n 
+0000000924 00000 n 
+0000001024 00000 n 
+0000001118 00000 n 
+0000001226 00000 n 
+0000001257 00000 n 
+trailer << /QTest 2 0 R /Root 1 0 R /Size 15 /ID [<31415926535897932384626433832795><31415926535897932384626433832795>] >>
+startxref
+1362
+%%EOF

qpdf/qtest/qpdf/copy-foreign-objects-out3.pdf

@@ -0,0 +1,92 @@
+%PDF-1.3
+%¿÷¢þ
+1 0 obj
+<< /Pages 3 0 R /Type /Catalog >>
+endobj
+2 0 obj
+<< /O1 4 0 R /O2 5 0 R /O3 6 0 R /This-is-QTest true >>
+endobj
+3 0 obj
+<< /Count 3 /Kids [ 7 0 R 8 0 R 6 0 R ] /Type /Pages >>
+endobj
+4 0 obj
+[ /This-is-O1 /potato << /O2 [ 3.14159 << /O2 5 0 R >> 2.17828 ] >> /salad /O2 5 0 R /Stream1 9 0 R ]
+endobj
+5 0 obj
+<< /K1 [ 2.236 /O1 4 0 R 1.732 ] /O1 4 0 R /This-is-O2 true >>
+endobj
+6 0 obj
+<< /Contents 10 0 R /MediaBox [ 0 0 612 792 ] /OtherPage 8 0 R /Parent 3 0 R /Resources << /Font << /F1 11 0 R >> /ProcSet [ /PDF /Text ] >> /Rotate 180 /This-is-O3 true /Type /Page >>
+endobj
+7 0 obj
+<< /Contents 12 0 R /MediaBox [ 0 0 612 792 ] /Parent 3 0 R /Resources << /Font << /F1 13 0 R >> /ProcSet 14 0 R >> /Type /Page >>
+endobj
+8 0 obj
+<< /Contents 15 0 R /MediaBox [ 0 0 612 792 ] /Parent 3 0 R /Resources << /Font << /F1 11 0 R >> /ProcSet [ /PDF /Text ] >> /Rotate 180 /This-is-O3-other-page true /Type /Page >>
+endobj
+9 0 obj
+<< /Stream2 16 0 R /This-is-Stream1 true /Length 18 >>
+stream
+This is stream 1.
+endstream
+endobj
+10 0 obj
+<< /Length 47 >>
+stream
+BT /F1 15 Tf 72 720 Td (Original page 2) Tj ET
+endstream
+endobj
+11 0 obj
+<< /BaseFont /Times-Roman /Encoding /WinAnsiEncoding /Subtype /Type1 /Type /Font >>
+endobj
+12 0 obj
+<< /Length 44 >>
+stream
+BT
+  /F1 24 Tf
+  72 720 Td
+  (Potato) Tj
+ET
+endstream
+endobj
+13 0 obj
+<< /BaseFont /Helvetica /Encoding /WinAnsiEncoding /Name /F1 /Subtype /Type1 /Type /Font >>
+endobj
+14 0 obj
+[ /PDF /Text ]
+endobj
+15 0 obj
+<< /Length 47 >>
+stream
+BT /F1 15 Tf 72 720 Td (Original page 3) Tj ET
+endstream
+endobj
+16 0 obj
+<< /Stream1 9 0 R /This-is-Stream2 true /Length 18 >>
+stream
+This is stream 2.
+endstream
+endobj
+xref
+0 17
+0000000000 65535 f 
+0000000015 00000 n 
+0000000064 00000 n 
+0000000135 00000 n 
+0000000206 00000 n 
+0000000323 00000 n 
+0000000401 00000 n 
+0000000601 00000 n 
+0000000747 00000 n 
+0000000941 00000 n 
+0000001046 00000 n 
+0000001143 00000 n 
+0000001243 00000 n 
+0000001337 00000 n 
+0000001445 00000 n 
+0000001476 00000 n 
+0000001573 00000 n 
+trailer << /QTest 2 0 R /Root 1 0 R /Size 17 /ID [<31415926535897932384626433832795><31415926535897932384626433832795>] >>
+startxref
+1678
+%%EOF

qpdf/test_driver.cc

@@ -916,6 +916,89 @@ void runtest(int n, char const* filename)
 	w.setStreamDataMode(qpdf_s_preserve);
 	w.write();
     }
+    else if (n == 25)
+    {
+        // The copy object tests are designed to work with a specific
+        // file.  Look at the test suite for the file, and look at the
+        // file for comments about the file's structure.
+
+        // Copy qtest without crossing page boundaries.  Should get O1
+        // and O2 and their streams but not O3 or any other pages.
+
+        QPDF newpdf;
+        newpdf.processFile("minimal.pdf");
+        QPDFObjectHandle qtest = pdf.getTrailer().getKey("/QTest");
+        newpdf.getTrailer().replaceKey(
+            "/QTest", newpdf.copyForeignObject(qtest));
+
+	QPDFWriter w(newpdf, "a.pdf");
+	w.setStaticID(true);
+	w.setStreamDataMode(qpdf_s_preserve);
+	w.write();
+    }
+    else if (n == 26)
+    {
+        // Copy the O3 page using addPage.  Copy qtest without
+        // crossing page boundaries.  In addition to previous results,
+        // should get page O3 but no other pages including the page
+        // that O3 points to.  Also, inherited object will have been
+        // pushed down and will be preserved.
+
+        QPDF newpdf;
+        newpdf.processFile("minimal.pdf");
+        QPDFObjectHandle qtest = pdf.getTrailer().getKey("/QTest");
+        QPDFObjectHandle O3 = qtest.getKey("/O3");
+        newpdf.addPage(O3, false);
+        newpdf.getTrailer().replaceKey(
+            "/QTest", newpdf.copyForeignObject(qtest));
+
+	QPDFWriter w(newpdf, "a.pdf");
+	w.setStaticID(true);
+	w.setStreamDataMode(qpdf_s_preserve);
+	w.write();
+    }
+    else if (n == 27)
+    {
+        // Copy O3 and the page O3 refers to before copying qtest.
+        // Should get qtest plus only the O3 page and the page that O3
+        // points to.  Inherited objects should be preserved.
+
+        QPDF newpdf;
+        newpdf.processFile("minimal.pdf");
+        QPDFObjectHandle qtest = pdf.getTrailer().getKey("/QTest");
+        QPDFObjectHandle O3 = qtest.getKey("/O3");
+        newpdf.addPage(O3.getKey("/OtherPage"), false);
+        newpdf.addPage(O3, false);
+        newpdf.getTrailer().replaceKey(
+            "/QTest", newpdf.copyForeignObject(qtest));
+
+	QPDFWriter w(newpdf, "a.pdf");
+	w.setStaticID(true);
+	w.setStreamDataMode(qpdf_s_preserve);
+	w.write();
+    }
+    else if (n == 28)
+    {
+        // Copy foreign object errors
+        try
+        {
+            pdf.copyForeignObject(pdf.getTrailer().getKey("/QTest"));
+            std::cout << "oops -- didn't throw" << std::endl;
+        }
+        catch (std::logic_error e)
+        {
+            std::cout << "logic error: " << e.what() << std::endl;
+        }
+        try
+        {
+            pdf.copyForeignObject(QPDFObjectHandle::newInteger(1));
+            std::cout << "oops -- didn't throw" << std::endl;
+        }
+        catch (std::logic_error e)
+        {
+            std::cout << "logic error: " << e.what() << std::endl;
+        }
+    }
     else
     {
 	throw std::runtime_error(std::string("invalid test ") +