Add QPDFObjectHandle::mergeDictionary()

Jay Berkenbilt
1 parent 104fd6da
Showing 8 changed files with 347 additions and 0 deletions
ChangeLog
include/qpdf/QPDFObjectHandle.hh
libqpdf/QPDFObjectHandle.cc
qpdf/qpdf.testcov
qpdf/qtest/qpdf.test
qpdf/qtest/qpdf/merge-dict.out
qpdf/qtest/qpdf/merge-dict.pdf
qpdf/test_driver.cc
 2018-12-31  Jay Berkenbilt  <ejb@ql.org>
+	* Add method QPDFObjectHandle::mergeDictionary(), which
+	recursively merges dictionaries with semantics designed for
+	merging resource dictionaries. See detailed description in
+	QPDFObjectHandle.hh.
+
 	* Add QPDFObjectHandle::Matrix, similar to
 	QPDFObjectHandle::Rectangle, as a convenience class for
 	six-element arrays that are used as matrices.
@@ -559,6 +559,28 @@ class QPDFObjectHandle
     QPDF_DLL
     bool isOrHasName(std::string const&);
+    // Merge dictionaries with the following behavior, where "object"
+    // refers to the object whose method is invoked, and "other"
+    // refers to the argument:
+    // * If either object or other is not a dictionary, do nothing
+    // * Otherwise
+    //   * For each key in other
+    //     * If key is absent in object, insert it
+    //     * If key is present in object
+    //       * If both values are dictionaries, merge the dictionary from
+    //         other into the one from object
+    //       * If both values are arrays, append scalar elements from
+    //         other's that are not present in object's onto object's,
+    //         and ignore non-scalar elements in other's
+    //       * Otherwise ignore
+    // The primary purpose of this method is to facilitate merging of
+    // resource dictionaries. Conflicts are ignored. If needed, a
+    // future version of qpdf may provide some mechanism for conflict
+    // resolution, such as providing a handler that is invoked with
+    // the path to the conflict.
+    QPDF_DLL
+    void mergeDictionary(QPDFObjectHandle other);
+
     // Return the QPDF object that owns an indirect object.  Returns
     // null for a direct object.
     QPDF_DLL
@@ -970,6 +992,10 @@ class QPDFObjectHandle
         ParserCallbacks* callbacks);
     std::vector<QPDFObjectHandle> arrayOrStreamToStreamArray(
         std::string const& description, std::string& all_description);
+    void mergeDictionaryInternal(
+        QPDFObjectHandle other,
+        std::set<QPDFObjGen>& visiting,
+        int depth);
     static void warn(QPDF*, QPDFExc const&);
     class Members
@@ -825,6 +825,109 @@ QPDFObjectHandle::isOrHasName(std::string const&amp; value)
     return false;
 }
+void
+QPDFObjectHandle::mergeDictionary(QPDFObjectHandle other)
+{
+    std::set<QPDFObjGen> visiting;
+    mergeDictionaryInternal(other, visiting, 0);
+}
+
+void
+QPDFObjectHandle::mergeDictionaryInternal(
+    QPDFObjectHandle other,
+    std::set<QPDFObjGen>& visiting,
+    int depth)
+{
+    if (depth > 100)
+    {
+        // Arbitrarily limit depth to avoid stack overflow
+        return;
+    }
+    if (! (isDictionary() && other.isDictionary()))
+    {
+        QTC::TC("qpdf", "QPDFObjectHandle merge top type mismatch");
+        return;
+    }
+    std::set<std::string> other_keys = other.getKeys();
+    for (std::set<std::string>::iterator iter = other_keys.begin();
+         iter != other_keys.end(); ++iter)
+    {
+        std::string const& key = *iter;
+        QPDFObjectHandle other_val = other.getKey(key);
+        if (hasKey(key))
+        {
+            QPDFObjectHandle this_val = getKey(key);
+            if (this_val.isDictionary() && other_val.isDictionary())
+            {
+                if (this_val.isIndirect() && other_val.isIndirect() &&
+                    (this_val.getObjGen() == other_val.getObjGen()))
+                {
+                    QTC::TC("qpdf", "QPDFObjectHandle merge equal indirect");
+                }
+                else if (this_val.isIndirect() &&
+                    (visiting.count(this_val.getObjGen())))
+                {
+                    QTC::TC("qpdf", "QPDFObjectHandle merge loop");
+                }
+                else
+                {
+                    QPDFObjGen loop;
+                    if (this_val.isIndirect())
+                    {
+                        loop = this_val.getObjGen();
+                        visiting.insert(loop);
+                        QTC::TC("qpdf", "QPDFObjectHandle merge shallow copy");
+                        this_val = this_val.shallowCopy();
+                        replaceKey(key, this_val);
+                    }
+                    QTC::TC("qpdf", "QPDFObjectHandle nested merge");
+                    this_val.mergeDictionaryInternal(
+                        other_val, visiting, 1 + depth);
+                    if (loop.getObj())
+                    {
+                        visiting.erase(loop);
+                    }
+                }
+            }
+            else if (this_val.isArray() && other_val.isArray())
+            {
+                std::set<std::string> scalars;
+                int n = this_val.getArrayNItems();
+                for (int i = 0; i < n; ++i)
+                {
+                    QPDFObjectHandle this_item = this_val.getArrayItem(i);
+                    if (this_item.isScalar())
+                    {
+                        scalars.insert(this_item.unparse());
+                    }
+                }
+                n = other_val.getArrayNItems();
+                for (int i = 0; i < n; ++i)
+                {
+                    QPDFObjectHandle other_item = other_val.getArrayItem(i);
+                    if (other_item.isScalar())
+                    {
+                        if (scalars.count(other_item.unparse()) == 0)
+                        {
+                            QTC::TC("qpdf", "QPDFObjectHandle merge array");
+                            this_val.appendItem(other_item);
+                        }
+                        else
+                        {
+                            QTC::TC("qpdf", "QPDFObjectHandle merge array dup");
+                        }
+                    }
+                }
+            }
+        }
+        else
+        {
+            QTC::TC("qpdf", "QPDFObjectHandle merge copy from other");
+            replaceKey(key, other_val);
+        }
+    }
+}
+
 // Indirect object accessors
 QPDF*
 QPDFObjectHandle::getOwningQPDF()
@@ -369,3 +369,11 @@ QPDFOutlineDocumentHelper string named dest 0
 QPDFOutlineObjectHelper loop 0
 qpdf required parameter 0
 qpdf required choices 0
+QPDFObjectHandle merge top type mismatch 0
+QPDFObjectHandle merge shallow copy 0
+QPDFObjectHandle nested merge 0
+QPDFObjectHandle merge array 0
+QPDFObjectHandle merge array dup 0
+QPDFObjectHandle merge copy from other 0
+QPDFObjectHandle merge loop 0
+QPDFObjectHandle merge equal indirect 0
@@ -895,6 +895,16 @@ $td-&gt;runtest(&quot;detect foreign object in write&quot;,
 show_ntests();
 # ----------
+$td->notify("--- Merge Dictionary ---");
+$n_tests += 1;
+
+$td->runtest("merge dictionary",
+             {$td->COMMAND => "test_driver 50 merge-dict.pdf"},
+             {$td->FILE => "merge-dict.out", $td->EXIT_STATUS => 0},
+             $td->NORMALIZE_NEWLINES);
+
+show_ntests();
+# ----------
 $td->notify("--- Parsing ---");
 $n_tests += 17;
+{
+  "/k1": "scalar1",
+  "/k2": 16059,
+  "/k3": {
+    "/a": "a",
+    "/b": "conflict: seen",
+    "/c": [
+      2,
+      3,
+      1
+    ],
+    "/d": {
+      "/x": 24,
+      "/y": 25,
+      "/z": 26
+    },
+    "/e": "e"
+  },
+  "/k4": {
+    "/A": 65,
+    "/B": 66,
+    "/C": 67,
+    "/indirect2": "8 0 R",
+    "/recursive": "9 0 R"
+  },
+  "/k5": [
+    "/one",
+    2,
+    "three",
+    [
+      "/four"
+    ],
+    "two"
+  ]
+}
+test 50 done
+%PDF-1.3
+%¿÷¢þ
+%QDF-1.0
+
+1 0 obj
+<<
+  /Pages 2 0 R
+  /Type /Catalog
+>>
+endobj
+
+2 0 obj
+<<
+  /Count 1
+  /Kids [
+    3 0 R
+  ]
+  /Type /Pages
+>>
+endobj
+
+%% Page 1
+3 0 obj
+<<
+  /Contents 4 0 R
+  /MediaBox [
+    0
+    0
+    612
+    792
+  ]
+  /Parent 2 0 R
+  /Resources <<
+    /Font <<
+      /F1 6 0 R
+    >>
+    /ProcSet 7 0 R
+  >>
+  /Type /Page
+>>
+endobj
+
+%% Contents for page 1
+4 0 obj
+<<
+  /Length 5 0 R
+>>
+stream
+BT
+  /F1 24 Tf
+  72 720 Td
+  (Potato) Tj
+ET
+endstream
+endobj
+
+5 0 obj
+44
+endobj
+
+6 0 obj
+<<
+  /BaseFont /Helvetica
+  /Encoding /WinAnsiEncoding
+  /Name /F1
+  /Subtype /Type1
+  /Type /Font
+>>
+endobj
+
+7 0 obj
+[
+  /PDF
+  /Text
+]
+endobj
+
+8 0 obj
+<<
+  /a (a)
+  /b (b)
+  /c [1 2]
+  /d << /x 24 /y (not seen) >>
+>>
+endobj
+
+9 0 obj
+<<
+  /A 65
+  /B 66
+  /indirect2 8 0 R
+  /recursive 9 0 R
+>>
+endobj
+
+xref
+0 10
+0000000000 65535 f 
+0000000025 00000 n 
+0000000079 00000 n 
+0000000161 00000 n 
+0000000376 00000 n 
+0000000475 00000 n 
+0000000494 00000 n 
+0000000612 00000 n 
+0000000647 00000 n 
+0000000729 00000 n 
+trailer <<
+  /Root 1 0 R
+  /Size 10
+  /ID [<f8c8da17f88e0dccac9f73ad9d0ee411><f8c8da17f88e0dccac9f73ad9d0ee411>]
+  /Dict1 <<
+    /k1 (scalar1)
+    /k3 <<
+      /b (conflict: seen)
+      /c [2 3]
+      /d << /y 25 /z 26 >>
+      /e (e)
+    >>
+    /k4 9 0 R
+    /k5 [
+      /one
+      2
+      (three)
+      [ /four ]
+    ]
+  >>
+  /Dict2 <<
+    /k1 (other: conflict: not seen)
+    /k2 16059
+    /k3 8 0 R
+    /k4 <<
+      /B (not seen)
+      /C 67
+      /indirect2 8 0 R
+      /recursive 8 0 R
+    >>
+    /k5 [
+      /one
+      (two)
+      << /six 6 >>
+      [ /five ]
+    ]
+  >>
+>>
+startxref
+805
+%%EOF
@@ -1754,6 +1754,17 @@ void runtest(int n, char const* filename1, char const* arg2)
             }
         }
     }
+    else if (n == 50)
+    {
+        // Test dictionary merge. This test is crafted to work with
+        // merge-dict.pdf
+        QPDFObjectHandle d1 = pdf.getTrailer().getKey("/Dict1");
+        QPDFObjectHandle d2 = pdf.getTrailer().getKey("/Dict2");
+        d1.mergeDictionary(d2);
+        std::cout << d1.getJSON().unparse() << std::endl;
+        // Top-level type mismatch
+        d1.mergeDictionary(d2.getKey("/k1"));
+    }
     else
     {
 	throw std::runtime_error(std::string("invalid test ") +